CN116193049A - Video processing method, device, computer equipment and storage medium - Google Patents

Abstract

The present application relates to a video processing method, device, computer equipment, storage medium and computer program product. The method acquires a video containing a target object; detects that the target object triggers a key event to automatically identify the wonderful moment of the target object, and determines the key segment in the video according to the key event, and determines the key from the video The related segment of the segment; the frame rate of the key segment can be adjusted to obtain the slow-motion video segment of the target frame rate; and the related segment can not be processed, and the adaptive processing of the frame rate can also be carried out, so that the related segment Synthesize the target video with the slow motion video clip. After the video is synthesized in this way, each segment of the target video has its own rhythm, and the center highlights the rhythm.

Description

Video processing method, device, computer equipment and storage medium

technical field

The present application relates to the technical field of image processing, in particular to a video processing method, device, computer equipment, storage medium and computer program product.

Background technique

Slow-motion shooting is a novel shooting method. At present, slow-motion videos are shot by people using electronic devices to manually capture the wonderful moments of the target object.

There are usually two ways to make slow-motion video: one is to use a fixed high frame rate for video shooting, and when saving the video, reduce the frame rate of the entire video to form slow-motion special effects, such as shooting at 120fps, the video will be If you save it at 15fps, you will get an 8x slow-motion video; the other is to use a video frame interpolation algorithm to synthesize a video with a higher frame rate to get slow-motion special effects. However, if the traditional slow-motion video production method is used to produce and process the video, the entire video will appear as a slow-motion video, resulting in a too single rhythm of the video.

Contents of the invention

Based on this, it is necessary to provide a video processing method, device, computer equipment, computer readable storage medium and computer program product for the above technical problems, which can identify and reduce the playback speed of key segments, so as to highlight the segments where key events occur, Add excitement to slow motion videos.

In a first aspect, the present application provides a video processing method. The methods include:

Obtain a video containing the target object;

Detecting that the target object triggers a key event;

determining a key segment in the video according to the key event, and determining related segments outside the key segment from the video;

Adjusting the frame rate of the key segment to obtain a slow-motion video segment of the target frame rate;

A target video is synthesized based on the relevant segments and the slow-motion video segments.

In one of the embodiments, the synthesizing the target video based on the related segment and the slow-motion video segment includes:

performing frame extraction processing on the relevant segment to obtain a sequence of video frames;

Synthesizing the sequence of video frames to obtain a video segment of the target frame rate;

and performing video synthesis on the video segment of the target frame rate and the slow-motion video segment to obtain a target video.

In one of the embodiments, the related segments include a first related segment and a second related segment;

The frame extraction processing of the relevant segments is performed to obtain a sequence of video frames, including:

performing frame extraction processing on the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence;

The synthesizing the sequence of video frames to obtain the video clips of the target frame rate includes:

Synthesizing the first video frame sequence and the second video frame sequence respectively to obtain the first video segment and the second video segment of the target frame rate;

The video synthesis of the synthesized video clip and the slow motion video clip includes:

performing video synthesis on the first video clip, the slow motion video clip and the second video clip.

In one of the embodiments, the frame extraction processing of the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence includes:

performing frame extraction processing on the first related segment and the second related segment according to the first frame sampling frequency to obtain a first video frame sequence and a second video frame sequence; or,

performing frame extraction processing on the first relevant segment and the second relevant segment according to a second frame extraction frequency to obtain a first video frame sequence and a second video frame sequence;

Wherein, the first frame drawing frequency is greater than the second frame drawing frequency, and the first frame drawing frequency is determined based on a ratio between the frame rate of the key segment and the target frame rate.

In one of the embodiments, the frame extraction processing of the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence includes:

performing frame extraction processing on the first relevant segment according to a first frame extraction frequency to obtain a first video frame sequence;

performing frame extraction processing on the second relevant segment according to a second frame extraction frequency to obtain a second video frame sequence;

Wherein, the first frame drawing frequency is greater than the second frame drawing frequency, and the first frame drawing frequency is determined based on a ratio between the frame rate of the key segment and the target frame rate.

In one of the embodiments, the frame extraction processing of the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence includes:

performing frame extraction processing on the first relevant segment according to a first frame extraction frequency to obtain a first video frame sequence;

According to each frame-drawing frequency in the frame-drawing frequency sequence, the sub-segments in the second relevant segment are sequentially subjected to frame-drawing processing, and different sub-segments are extracted for each frame-drawing frequency during the frame-drawing process to obtain a second video frame sequence;

Wherein, each frame drawing frequency in the frame drawing frequency sequence is increasing or decreasing.

In one of the embodiments, the frame rate of the key segment is adjusted to obtain a slow-motion video segment of the target frame rate, including:

determining a target frame rate, the target frame rate being lower than the frame rate of the key segment;

converting the frame rate of the key segment to the target frame rate to obtain a slow-motion video segment.

In one of the embodiments, the determining the key segment in the video according to the key event, and determining the related segments other than the key segment from the video include:

Determining the critical time period during which the key event exists in the video;

Screening key segments in the video according to the key time period;

determining at least one relevant time period in the preceding preset time period of the critical time period, the subsequent preset time period of the critical time period, and the interval time period between each of the critical time periods;

Find relevant segments in the video according to the relevant time period.

In one of the embodiments, the detecting the target object to trigger a key event includes:

Obtain scene information of the video;

Detecting the action of the target object according to the scene information;

If it is determined that the action of the target object belongs to the preset action of the scene information, then determine the key event that the target object triggers the preset action.

In one of the embodiments, the detection of the action of the target object according to the scene information includes:

If the scene information represents a basketball scene, then determine that the preset action of the scene information is a shooting action;

According to the human body posture corresponding to the basketball scene information, extracting the posture key points of the target object from the video;

identifying whether the action of the target object belongs to a shooting action according to the key points of the posture;

The action of determining the target object belongs to the preset action of the scene information, including:

The action of determining the target object belongs to the shooting action.

In one of the embodiments, the detecting the target object to trigger a key event includes:

determining a video frame of the video;

In the video picture, detect the area parameter of the target object; the area parameter includes the area of the target object in the video picture, or the ratio between the area of the target object and the area of the video picture;

When the area parameter is greater than a preset area parameter threshold, it is determined that the target object triggers a key event.

In a second aspect, the present application also provides a video processing device, the device comprising:

The video acquisition module is used to acquire the video containing the target object;

An event detection module, configured to detect that the target object triggers a key event;

Segment extraction module, for determining the key segment in the video according to the key event, and determining the relevant segments outside the key segment from the video;

A key segment speed regulation module, used to adjust the frame rate of the key segment to obtain a slow-motion video segment of the target frame rate;

A video synthesis module, configured to synthesize a target video based on the relevant segments and the slow-motion video segments.

In a third aspect, the present application also provides a computer device. The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the steps of video processing in any of the above embodiments when executing the computer program.

In a fourth aspect, the present application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the steps of video processing in any of the above-mentioned embodiments are implemented.

In a fifth aspect, the present application also provides a computer program product. said computer program product,

It includes a computer program, and when the computer program is executed by a processor, the steps of video processing in any of the above-mentioned embodiments are realized.

The above-mentioned video processing method, device, computer equipment, storage medium and computer program product acquire a video containing a target object; detect that the target object triggers a key event to automatically identify the wonderful moment of the target object, and determine the video according to the key event key clips in the , and from said video

Determine the relevant segments outside the key segment; the frame rate can be adjusted to the key segment to obtain a slow-motion video segment with a target frame rate of 0; and the relevant segment can not be processed, and the frame rate can also be adaptively processed, so that The relevant segments and the slow-motion video segments are combined into a target video. Each segment of the target video synthesized in this way has its own rhythm, and the center highlights the rhythm clearly.

Description of drawings

Fig. 1 is an application environment diagram of a video processing method in an embodiment;

Fig. 2 is a schematic flow chart of a video processing method in an embodiment;

Fig. 3 is a schematic flow chart of relevant segment frame extraction in an embodiment;

Fig. 4 is a structural block diagram of a video processing device in an embodiment;

Figure 5 is an internal block diagram of a computer device in one embodiment.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, and are not intended to limit the present application.

5. The video processing method provided in the embodiment of the present application may be applied to the application environment shown in FIG. 1 .

Wherein, the terminal 102 communicates with the server 104 through the network. The data storage system can store data that needs to be processed by the server 104 . The data storage system can be integrated on the server 104, or placed on the cloud or other network servers.

Among them, the terminal 102 can be various shooting devices, and specific shooting devices include but are not limited to various cameras, panoramic cameras, sports cameras, personal computers, notebook computers, smart phones, tablet computers, Internet of Things devices and portable wearable devices , IoT devices can be smart TVs, smart car devices, etc. Portable wearable devices can be smart watches, headsets, and the like. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers. This solution can be implemented based on the terminal 102 , can also be implemented based on the server 104 , and can also be implemented through the interaction between the terminal 102 and the server 104 .

In one embodiment, as shown in FIG. 2, a video processing method is provided, and the method is applied to the server 104 in FIG. 1 as an example for illustration, including the following steps:

Step 202, acquire the video containing the target object.

The target object is a prominent object in the video; the target object can be a person that exists in reality, the object can be a virtual character, or an object or an animal; typical target objects include but are not limited to people, human heads, human Faces, cats, dogs, vehicles. For the video containing the target object, you can choose a high frame rate video, and use a terminal with a pan/tilt or other parameters that can adjust image acquisition parameters. Exemplarily, the high frame rate may be 120fps or 240fps, or other higher than 24fps or 30fps commonly used in ordinary video.

The process of obtaining a video containing the target object can use a target detection algorithm; the target detection algorithm can be a common detection method, and the common detection method can be one or more detection methods based on manual features, and the detection method based on manual features Including template matching method, key point matching method, and key feature method; common detection methods can also be one or more detection methods based on convolutional neural network technology, and convolutional neural network technology can use YOLO (You Only Look Once, You Only look once), SSD (Single Shot MultiBox, single-shot multi-box), R-CNN (Region-based Convolutional Neural Networks, region-based convolutional neural network) or Mask R-CNN (MaskRegion-based Convolutional Neural Networks, with mask One or more models in a region-based convolutional neural network for membranes. Exemplarily, the YOLO model may be one or more models of YOLOv1-YOLOv5, YOLOR, YOLOX and other versions.

Optionally, when the shooting instruction of the target object is received, each object in the video picture is identified, and the reference point of the region of interest (such as the center point of the rectangular frame) and the video picture reference point (such as the center point of the video picture) of each object are determined. The distance between , and the reference point of the region of interest with the smallest distance is determined as the target object. When the shooting instruction of the target object is received, each object display in the video frame is recognized and displayed, so as to determine the target object based on the user's selection instruction. The screen displaying the video image may be a built-in screen of an electronic device (mobile phone, camera), or a computer screen displaying a PC client. Exemplarily, a user initiates

Object tracking can be initiated from the PC client (use mouse, keyboard operation), and directly operate electronic equipment (press button 5, touch screen); when the target object includes at least one of human beings, human heads and human faces, it can be

Determined by preset gestures or body poses used by the subject in front of the camera. In addition, object tracking can also be automatically initiated by preset specific events: for example, automatically initiated within 10 seconds after powering on; for example, automatically initiated when an object enters a specified area in the screen (such as an area with the size of NxN pixels in the center of the screen).

When the video containing the object is an image to be collected, the image is collected in real time for the position change of a certain target object. This image collection process may be called object tracking or person tracking.

When performing object tracking, an arbitrary frame image is used as a first video frame, and a preset target tracking algorithm is used to perform target tracking to obtain a target rectangular frame of the first video frame. It can be a discriminative correlation filter (Discriminative Correlation Filter, DCF) and other related filter class trackers, which can be

Siamese Region Proposal Network (SiamRPN) and other trackers based on CNN technology can also be other trackers.

In one embodiment, during the video capture process, the capture parameters of the video frame are adjusted according to the position change of the target object in the video, and the target object is controlled to be located at a preset position of the video frame according to the capture parameters. Thus, the video includes a plurality of video frames, and the video frames are determined according to changes in the position of the target object.

and the target object is located at a preset position of the video frame, so as to change the image acquisition parameters of the video frame along with the movement of the target object.

In a feasible embodiment, before detecting that a target object triggers a key event, the method includes:

In the first video frame of the video, detect the salient area and the associated area corresponding to the salient area; calculate the difference between the salient area and the associated area; determine the target composition information according to the difference; according to the target composition information,

Updating the image acquisition parameters of the first video frame; performing image acquisition according to the updated image acquisition parameters to obtain the second video frame; detecting the target object in the second video frame.

Wherein, a salient area refers to an image area determined in the first video frame according to a certain salient feature, and an associated area is an area determined by association according to a certain relationship between the salient area and the associated area.

The difference between the salient area and the associated area, the difference between the salient area and the associated area can be height, width, area difference value or other difference value, the difference value includes the ratio or difference, the difference between the salient area and the associated area is used to determine the corresponding pattern of the composition The target composition information of the target composition information is used to determine the distribution information of the salient area and the associated area, and then the first video frame is adjusted through the image acquisition parameters. Taking the composition mode used to track the target object as an example, moving the gimbal according to the updated image acquisition parameters of the target composition information can make the target object in the center of the screen or near the center of the screen, and can also make the target object in a position above the center of the screen , so as to change the composition with the position of the head and face.

When there is no need to control the target object at the preset position of the video screen, the end operation of object tracking can be performed. Among them, the target circumscribed rectangular frame can be displayed in real time on the user interface, and an X (close) icon is attached to the upper right corner at the same time. When the user clicks the X icon, the tracking ends; when the user clicks an area other than the target in the screen, the tracking ends; The tracking can be ended when the user clicks a preset button.

Step 204, detecting that the target object triggers a key event.

A key event is a certain action, position, or posture of the target object. An event that will be triggered when the expression, area or other information meets the key event conditions. Key events can be defined according to specific scenarios.

In a possible embodiment, detecting a target object to trigger a key event includes: obtaining scene information of the video; detecting the action of the target object according to the scene information; A key event for an object to trigger a preset action.

The scene information represents the scene where the target object is located, and each scene has its own corresponding preset action to improve the accuracy of key event recognition. The scene information can be scenes such as sports scenes or teaching scenes, or it can be a refined scene obtained by further subdividing these coarse-grained scenes, and the accuracy of key event recognition of the refined scene can be further improved.

Taking the sports scene as an example, the preset actions of the refined scene include but are not limited to: head shaking, turning back, body wave and other preset actions in the dance scene, and preset actions such as vacating in the skiing and skateboarding scene , Preset actions such as shooting, layup, dunk and so on in the basketball scene. Among them, the video model specifically refers to a deep learning model that takes multi-frame video as input, such as expanded three-dimensional space-time domain network (Expand 3D, X3D), fast and slow network (Slowfast) and other models.

In an exemplary embodiment, detecting the action of the target object in the video according to the scene information includes: when the video scene is a basketball scene, identifying the action of the target object based on a video frame sequence.

Correspondingly, the action of determining the target object corresponds to the preset action of the scene information, including: the action of determining the target object is any jumping action of one foot in the air, both feet in the air, or both feet in the air.

In a possible embodiment, detecting the action of the target object according to the scene information includes: if the scene information represents a basketball scene, then determining that the preset action of the scene information is a shooting action; according to the human body posture corresponding to the basketball scene, extracting The posture key points of the target object; according to the posture key points, it is recognized whether the action of the target object belongs to the shooting action.

Correspondingly, the action of determining the target object belongs to the preset action of the scene information, including: the action of determining the target object belongs to the shooting action.

Optionally, the pose key points may be pose extraction templates, which include but are not limited to any type of pose key points such as shape edge points, muscle change points, and bone key points of various parts of the human body. User actions can be planned through various types of gesture key points to improve the accuracy of recognition.

Therefore, in the basketball scene, the human body key points of the target object are estimated by applying the human body pose based on deep learning, and the human body pose is estimated through a relatively low amount of data, so that the parallel conditions of detecting key events and video extraction are reduced, so that the steps S204 is executed synchronously with other steps; and then the shooting action is recognized according to the key points of the human body, so as to use the highlight moment in the basketball scene as a slow-motion video clip.

In a possible embodiment, detecting the target object to trigger a key event includes: detecting the expression of the target object based on CNN or other classification networks; when it is determined that the expression of the target object belongs to the preset expression set, according to the expression of the preset expression set There is a time period to determine the target object triggering the key event.

In a possible embodiment, detecting a target object to trigger a key event includes: determining a video frame of a video; in the video frame, detecting an area parameter of the target object; the area parameter includes the area of the target object in the video frame, or the target object The ratio between the area and the area of the video frame; when the area parameter is greater than the preset area parameter threshold, it is determined that the target object triggers a key event.

If the area parameter is the area of the target object in the video screen, and the area of the target object is the area of the target object in the video screen, then the preset area parameter threshold is also the threshold of the area; if the area parameter is the area ratio between the target object area and the video screen , then the preset area parameter threshold is also the threshold of the area ratio.

Thus, the distance between the target object and the terminal is determined based on the area parameter, and whether the distance indicates that the target object is close enough to the terminal is determined based on the preset area parameter threshold, and the video segment at this time is determined as a key segment.

In a possible embodiment, detecting that a target object triggers a key event includes: providing an operation option on a user interface of the preview video to receive a user instruction, and determining a key event touch point according to the received user instruction.

Step 206, determine the key segment in the video according to the key event, and determine the relevant segments other than the key segment from the video.

Both the key clip and its related clips belong to the video clips in the original video. The key segment is a video segment obtained by determining the key segment ID by the key event, and then mapping based on the key segment ID; the key segment can determine the related segment ID, and then map based on the related segment ID, to obtain the related segment of the key segment. When the playback speed of a key segment changes, the related segments of the key segment will be adaptively adjusted, so that the smoothness of the target video or other video evaluation data fluctuate less, ensuring video quality.

There is a corresponding relationship between the time period in which the key event exists and the key fragment identifier, and the key fragment identifier can be mapped based on the time period in which the key event exists, or the time period in which the key event exists can be directly used as the key fragment identifier.

The related segment ID of the key segment is associated with the key segment ID; and the related segment ID of the key segment can also be an ID or a timestamp. The relevant segment identifier may be a time period before and after the key segment, so as to determine a video segment in a time period before and after the key segment as a relevant segment.

In one embodiment, determining the key segment in the video according to the key event, and determining the relevant segment outside the key segment from the video, includes: determining the key time period in which the key event exists in the video; screening the key segment in the video according to the key time segment Segment; in the previous preset time period of the key time period, the next preset time period of the key time period, and the interval time period between each key time period, determine at least one relevant time period; search for the video according to the relevant time period related fragments.

The key time period is a key segment identification, and the key segment can be identified by the time stamp of the video frame. The previous preset time period of the key time period, the subsequent preset time period of the key time period, and the interval time period between each key time period can be used as the input data in the association relationship, and the input data in the association relationship The relevant time period is determined, and then the relevant part of the key part is calculated according to the relevant time period. Optionally, the time stamp in the relevant time period may be used as the identifier of each video frame in the relevant segment, or the identifier of each video frame in the relevant segment may be determined based on the result obtained from the time stamp mapping in the relevant time period.

Step 208, adjusting the frame rate of the key segment to obtain a slow-motion video segment with a target frame rate.

In one embodiment, adjusting the frame rate of the key segment to obtain the slow-motion video segment of the target frame rate includes: determining the target frame rate, the target frame rate being lower than the frame rate of the key segment; The frame rate of the key segment is converted to the target frame rate to obtain a slow motion video segment.

In an optional embodiment, converting the frame rate of the key segment into the target frame rate to obtain the slow-motion video segment includes: acquiring images in the key segment, and synthesizing images in the key segment according to the target frame rate image to get a slow motion video clip.

The target frame rate is lower than the frame rate of the key segment, so the number of image frames per second of the key segment is much greater than the number of image frames per second required by the slow-motion video segment, so that the images in the key segment form a slow-motion video segment, Realize slow motion special effects. Exemplarily, the frame rate of the key segment is 240FPS, and the frame rate of the slow-motion video is 30FPS, so as to reduce the speed of the key segment to about one-eighth of the playback speed of the original video.

Since the number of frames of the image in the key segment is fixed, and the number of image frames per second of the key segment is much greater than the number of image frames per second required by the slow-motion video segment, the duration of the slow-motion video segment exceeds the duration of the key segment. To form slow-motion video clips, so as to clearly show the movement of the target object in key clips, and then form a sharp contrast with related clips to improve the sense of rhythm.

In one embodiment, converting the frame rate of the key segment into the target frame rate to obtain the slow-motion video segment includes: determining the first slow-motion video segment of the first target frame rate; Determine the motion data of the adjacent video frame at the middle moment of the adjacent video frame in the segment; filter the corresponding event image at the intermediate moment of the adjacent video frame from multiple event images of the motion data of the key segment; according to the adjacent video frame and its corresponding event image to obtain the relative motion state at the middle moment of the corresponding event image; according to the relative motion state, the video frame interpolation processing of the intermediate frame is performed on the corresponding event image to obtain the second slow-motion video segment, and the second slow-motion video segment has a relatively High frame rate; wherein, the frame rate of the second slow-motion video segment is greater than that of the first slow-motion video segment. Thus, the frame rate of the first slow-motion video clip can be increased again to obtain the second slow-motion video clip, so as to increase the smoothness of the target video.

Step 210, synthesizing a target video based on the related segment and the slow-motion video segment.

In an optional embodiment, synthesizing the target video based on related segments and slow-motion video segments includes: performing frame extraction processing on the related segments to obtain a sequence of video frames; synthesizing the sequence of video frames to obtain the target frame video clips at the target frame rate; performing video synthesis on the video clips at the target frame rate and the slow-motion video clips to obtain the target video.

The frame extraction process for related segments can select video frames based on the frame extraction frequency, filter video frames according to the frame extraction interval, or use other elements, as long as the element can extract video frame sequences from related segments. A sequence of video frames is an ordered sequence of video frames extracted from related segments. The sequence of each video frame may be a time sequence, or a sequence defined by a user or other programs.

In one application scenario, the frame rate is used to keep the playback speed of the target frame rate video segment and related segments unchanged. When the ratio between the frame rate of the key segment and the target frame rate is 8, a certain frame extraction frequency is 0.125, that is, according to the timestamp or other sorting elements, one frame of image is sequentially extracted from every 8 frames of images in the relevant segment as Images in the sequence of video frames, and so on until a sequence of video frames is obtained. When the ratio between the frame rate of the key segment and the target frame rate is 10, a certain frame extraction frequency is 0.1, that is, according to the timestamp or other sorting elements, every 10 frames of images in the relevant segment are sequentially extracted. By analogy, until a sequence of video frames is obtained.

The frame extraction interval is used as an equivalent replacement for the frame extraction frequency. When the ratio between the frame rate of the key segment and the target frame rate is 8, then every 7 frames of images in the relevant segment are used as the frame extraction interval, and the 8th frame is selected according to the frame extraction interval. Frame images as images in a sequence of video frames, and so on.

In one embodiment, performing frame extraction processing on the related segment to obtain a video frame sequence includes: when there is a related segment, extracting a video frame sequence from the related segment.

Correspondingly, synthesizing the sequence of video frames to obtain the video segment at the target frame rate includes: sequentially combining the sequence of video frames extracted from the related segments to obtain the video segment at the target frame rate.

In one embodiment, performing frame extraction processing on the related segments to obtain a video frame sequence includes: when there are two related segments, extracting a video frame sequence of each related segment from each related segment.

Correspondingly, synthesizing the sequence of video frames to obtain a video segment at the target frame rate includes: sequentially combining the sequence of video frames extracted from the same related segment to obtain a video segment at the target frame rate of the two related segments.

In one embodiment, performing video synthesis on the video clip of the target frame rate and the slow motion video clip includes: if there is a video clip of the target frame rate, combining the video clip of the target frame rate with the slow motion Video clip stitching; if there are two video clips with a target frame rate, the two video clips with a target frame rate are spliced with the slow-motion video clip.

Therefore, by means of frame extraction, the relevant clips are converted into video clips with a target frame rate, and the video clips with a target frame rate and the slow-motion video with a target frame rate have a matching frame rate, which can ensure the smooth connection process between the two better.

In one embodiment, the related segment includes a first related segment and a second related segment;

The performing frame extraction processing on the relevant segment to obtain a video frame sequence includes: performing frame extraction processing on the first relevant segment and the second relevant segment to obtain a first video frame sequence and a second video frame sequence ; Wherein, the frame extraction processing of the first video frame sequence and the second video frame sequence may be the same or different, so that the first video frame sequence and the second video frame sequence may have their own characteristics. Exemplarily, if the first video frame sequence is obtained by frame extraction according to the first frame extraction frequency or the first frame extraction interval, the second related segment may be extracted according to the second frame extraction frequency or the second frame extraction interval obtained by frame processing, and the second related segment may also be obtained by mapping.

The synthesizing the video frame sequence to obtain the video segment at the target frame rate includes: synthesizing the first video frame sequence and the second video frame sequence respectively to obtain the first video segment and the second video segment at the target frame rate Two video clips; wherein, the sorting manners of the first video frame sequence and the second video frame sequence can be the same or different, and the synthesis methods of the first video frame sequence and the second video frame sequence can be the same or different, to The first video segment and the second video segment of the target frame rate may have respective advantages.

In an optional embodiment, if the first video segment is obtained by sequentially synthesizing the video frames in the sequence according to the time stamp, the second video segment may be obtained by synthesizing the video frames in the first video frame sequence in sequence according to the time stamp frame, and the second video segment may be obtained by sequentially synthesizing each video frame in the second video frame sequence according to the reverse order of the time stamp. If the first video clip is obtained by splicing each video frame in the first video clip, then the second video clip can be obtained by splicing each video frame in the second video clip, and the second video clip It may also be obtained by superimposing each video frame in the second video segment according to the position matching results in each video frame.

The performing video synthesis on the synthesized video segment and the slow motion video segment includes: performing video synthesis on the first video segment, the slow motion video segment and the second video segment. Wherein, the composite processing of the first video clip, the slow motion video clip, and the composite processing of the slow motion video clip and the second video clip can be the same or different means, so that each clip of the target video has respective advantages.

Firstly, the granularity of the related segments is refined into the first related segment and the second related segment, then the first video frame sequence is extracted from the first related segment, and the second video frame sequence is extracted from the second related segment; then based on the first video frame Sequentially compositing a first video clip at a target frame rate; and then performing video compositing on the first video clip, the slow-motion video clip, and the second video clip. In this way, the needs of product managers can be further satisfied, and the details of each part of the target video can be optimized separately.

In the above video processing method, firstly, the target object is tracked during shooting, so as to control the target object to be near the center of the picture, realize object tracking, and make the main body stand out; secondly, the key segment is identified, and the key segment and related segments are used With different double speeds, the key events are highlighted, and the sense of rhythm is improved, so that the rhythm is fast and slow, and it has a strong visual shock, and the degree of excitement is far superior to the existing methods.

In one embodiment, the related segments include a first related segment and a second related segment; the target video is synthesized based on the related segment and the slow-motion video segment, including:

Step 302: Perform frame extraction processing on the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence.

In a possible embodiment, the frame extraction processing is performed on the first related segment and the second related segment to obtain the first video frame sequence and the second video frame sequence, including: according to the first frame extraction frequency, the first related segment Perform frame extraction processing with the second related segment to obtain the first video frame sequence and the second video frame sequence; or, according to the second frame extraction frequency, perform frame extraction processing on the first related segment and the second related segment to obtain the first video frame sequence A video frame sequence and a second video frame sequence; wherein, the first frame extraction frequency is greater than the second frame extraction frequency, and the first frame extraction frequency is determined based on the ratio between the frame rate of the key segment and the target frame rate.

The first frame rate is determined based on the ratio between the frame rate of the key segment and the target frame rate. Therefore, the video frame sequence obtained according to the first frame rate can be easily synthesized into a video segment, which can be played at the same speed as the original video. Make the first video clip and the second video clip both ordinary original speed video clips; and because the first frame extraction frequency is greater than the second frame extraction frequency, the video frame sequence obtained according to the second extraction frame frequency can easily synthesize the original speed video The segments can be played at a speed greater than the original video, so that both the first video segment and the second video segment are accelerated video segments with a preset multiple.

It is used to determine the ratio between the frame rate of the key segment and the target frame rate. When the frame rate of the key segment and the related segment are the same, the ratio between the frame rate of the video and the target frame rate can be directly used as the preset frame rate ratio . The preset frame rate ratio can be directly used as the original frame rate, or the preset frame rate ratio can be mapped or matched to obtain the original frame rate.

Thus, the first frame rate is determined based on the ratio between the frame rate of the key segment and the target frame rate, and the frame is drawn according to one of the first frame rate and the second frame rate, so that there is more Well control the number of video frames in the first video frame sequence and the second video frame sequence, so that the first related video and the second related video are simultaneously converted into common original speed video clips, or the first related video and the second related video Relevant videos are simultaneously converted into accelerated video clips with preset multiples, so that the subject of the target video is highlighted, the key events are highlighted, the rhythm is fast and slow, and it has a strong visual shock. Thus, a target video of "original speed-slow speed-original speed" or "acceleration-slow speed-acceleration" can be formed.

In a possible embodiment, performing frame extraction processing on the first related segment and the second related segment to obtain the first video frame sequence and the second video frame sequence includes: performing processing on the first related segment according to the first frame extraction frequency Frame extraction processing to obtain the first video frame sequence; perform frame extraction processing on the second relevant segment according to the second frame extraction frequency to obtain the second video frame sequence; wherein the first frame extraction frequency is greater than the second frame extraction frequency, and the second frame extraction frequency is greater than the second frame extraction frequency. A frame rate is determined based on the ratio between the frame rate of the key segment and the target frame rate.

The first frame sampling frequency is greater than the second frame sampling frequency, so when performing frame sampling processing on related segments of the same duration, the number of video frames extracted according to the first frame sampling frequency is greater than the number of video frames extracted according to the second frame sampling frequency. Therefore, when the first video frame sequence is determined according to the first frame extraction frequency, and the first video segment synthesized according to the first video frame sequence is played at the same speed as the original video, the second video frame sequence is determined according to the second frame extraction frequency , and the second video segment synthesized according to the second video frame sequence has a faster playback speed. Exemplarily, when the first frame extraction frequency is 0.125, the second frame extraction frequency can be 0.0625, that is, every 8 frames of images in the first relevant segment extract an image, so that the first video segment maintains the original speed; the second One frame of images is extracted from every 16 frames of images in the relevant segment, so that the second video segment is twice the speed of the original video.

Therefore, the first related segment and the second related segment are processed separately by using different frame extraction frequencies, so that one of the related segments is a normal original speed video segment, and the other segment is an accelerated video segment with a preset multiple, so as to adapt to the corresponding products. Therefore, the target video of "acceleration-slow speed-original speed" and "original speed-slow speed-acceleration" can be formed.

In a possible embodiment, performing frame extraction processing on the first related segment and the second related segment to obtain the first video frame sequence and the second video frame sequence includes: performing processing on the first related segment according to the first frame extraction frequency Frame extraction processing to obtain the first video frame sequence; according to each frame extraction frequency in the frame extraction frequency sequence, frame extraction processing is performed on the sub-segments in the second relevant segment in sequence, and each frame extraction frequency is extracted during the frame extraction process The second video frame sequence is obtained for different sub-segments; wherein, the first frame extraction frequency is greater than each frame extraction frequency in the frame extraction frequency sequence; each frame extraction frequency in the frame extraction frequency sequence is increasing or decreasing.

Each frame frequency in the frame frequency sequence can be dynamically generated based on some functions, or it can be preset, as long as each frame frequency in the frame frequency sequence is in the order of increasing or decreasing frame frequency arranged, and used to perform frame extraction processing on each sub-segment in each second related segment. Each sub-segment in the second related segment has the same number of video frames, and each sub-segment can determine the position of the sub-segment according to the sequence of video identification such as the time stamp of each sub-segment. The positions of the sub-segments may be determined as positions of each sub-segment relative to the second related segment, or may be determined as positions of each sub-segment relative to the original video. Wherein, the sequence of the frame sampling frequency and the sequence of the video identification are both positively correlated with the time stamp.

When each frame sampling frequency is arranged in increasing order, the more video frames are extracted from each sub-segment by each frame sampling frequency, the longer the duration of the second video segment is, then when the range of action changes remains the same, the second The picture of the video segment presents the effect of gradually decelerating; when each frame sampling frequency is arranged in descending order, the less the number of video frames extracted by each frame sampling frequency from the sub-segment, the shorter the duration of the second video segment, then When the change range of the motion remains unchanged, the picture of the second video segment presents the effect of gradually accelerating. As a result, "progressive acceleration-slow motion-original speed", "gradual deceleration-slow motion-original speed", "original speed-slow motion-progressive acceleration", and "original speed-slow motion-progressive deceleration" etc. model.

Step 304, synthesizing the first video frame sequence and the second video frame sequence respectively to obtain the first video segment and the second video segment at the target frame rate.

Step 306, performing video synthesis on the first video segment, the slow motion video segment and the second video segment.

As a result, the key fragments are identified, different double speeds are used for key fragments and related fragments, key events are highlighted, and related fragments are controlled to play at original speed, multiple speed or gradual speed, so as to better control The sense of rhythm further enhances the visual shock of the target video.

In an exemplary embodiment, step 210 of this application is described by taking the case of generating the first video frame sequence and the second video frame sequence at the first frame rate as an example.

Record a video with a frame rate of 240fps, and get a 30-second video, with a total of 240*30=7200 frames. In the 7200 frame, the terminal automatically recognizes the action of jumping into the air, and determines that the 10-11 second interval (the 2400th frame to the 2639th frame, a total of 240 frames) is a key segment, and the target object in the key segment is a humanoid object playing a skateboard; Seconds 0-10 (frame 0 to frame 2399) of the original video, and seconds 11-30 (frame 2639 to frame 7200) of the original video are sequentially used as the first relevant segment and the second relevant segment. When synthesizing a 240fps video into a 30fps video, modify the 10-11 second interval of the original video to 8 times slow motion (from 1 second to 8 seconds), and use the normal speed for other parts to get a 30fps video with a length of 37 seconds.

Specifically, the video segment of the 0-10th second of the original video is determined as the first relevant segment, and is processed with the first frame sampling frequency. It reserves 1 frame every 8 frames (each sub-segment), that is, retains frame 0 and discards frames 1-7, retains frame 8 and discards frames 9-15, and so on, and stores at 30fps. After this step, this part of the 10-second video at 240fps becomes the first video clip at 30fps at the original speed of 10 seconds.

The video segment of the 10th-11th second of the original video is determined as the key segment, and all images of the key segment are stored at 30fps. After this step, the 1-second key clip at 240fps becomes an 8-second slow-motion video clip at 30fps.

Determine the video segment of the 11th-30th second of the original video as the second relevant segment, and process it with the second frame sampling frequency. It reserves 1 frame every 8 frames (each sub-segment), that is, retains frame 0 and discards frames 1-7, retains frame 8 and discards frames 9-15, and so on, and stores at 30fps. After this step, the part of the 19-second video at 240fps becomes the second video clip at the original speed of 19 seconds at 30fps.

According to the timestamp, sequentially synthesize the first video segment, the slow-motion video segment, and the second video segment to obtain the target video.

It should be understood that although the steps in the flow charts involved in the above embodiments are shown sequentially according to the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least some of the steps in the flow charts involved in the above-mentioned embodiments may include multiple steps or stages, and these steps or stages are not necessarily executed at the same time, but may be performed at different times For execution, the execution order of these steps or stages is not necessarily performed sequentially, but may be executed in turn or alternately with other steps or at least a part of steps or stages in other steps.

Based on the same inventive concept, an embodiment of the present application further provides a video processing device for implementing the above-mentioned video processing method. The solution to the problem provided by the device is similar to the implementation described in the above method, so the specific limitations in one or more video processing device embodiments provided below can refer to the above definition of the video processing method, I won't repeat them here.

In one embodiment, as shown in Figure 4, a video processing device is provided, including:

A video acquisition module 402, configured to acquire a video containing a target object;

An event detection module 404, configured to detect that the target object triggers a key event;

Segment extraction module 406, for determining the key segment in the video according to the key event, and determining the relevant segments outside the key segment from the video;

The key segment speed adjustment module 408 is used to adjust the frame rate of the key segment to obtain the slow-motion video segment of the target frame rate;

A video synthesis module 410, configured to synthesize a target video based on the related segments and the slow-motion video segments.

In one of the embodiments, the video synthesis module 410 is configured to:

performing frame extraction processing on the relevant segment to obtain a sequence of video frames;

Synthesizing the sequence of video frames to obtain a video segment of the target frame rate;

and performing video synthesis on the video segment of the target frame rate and the slow-motion video segment to obtain a target video.

In one of the embodiments, the related segments include a first related segment and a second related segment;

The video synthesis module 410 is used for:

performing frame extraction processing on the first related segment and the second related segment to obtain a first video frame sequence and a second video frame sequence;

Synthesizing the first video frame sequence and the second video frame sequence respectively to obtain the first video segment and the second video segment of the target frame rate;

performing video synthesis on the first video clip, the slow motion video clip and the second video clip.

In one of the embodiments, the video synthesis module 410 is configured to:

performing frame extraction processing on the first related segment and the second related segment according to the first frame sampling frequency to obtain a first video frame sequence and a second video frame sequence; or,

performing frame extraction processing on the first relevant segment and the second relevant segment according to a second frame extraction frequency to obtain a first video frame sequence and a second video frame sequence;

Wherein, the first frame drawing frequency is greater than the second frame drawing frequency, and the first frame drawing frequency is determined based on a ratio between the frame rate of the key segment and the target frame rate.

In one of the embodiments, the video synthesis module 410 is configured to:

performing frame extraction processing on the first relevant segment according to a first frame extraction frequency to obtain a first video frame sequence;

performing frame extraction processing on the second relevant segment according to a second frame extraction frequency to obtain a second video frame sequence;

Wherein, the first frame drawing frequency is greater than the second frame drawing frequency, and the first frame drawing frequency is determined based on a ratio between the frame rate of the key segment and the target frame rate.

In one of the embodiments, the video synthesis module 410 is configured to:

performing frame extraction processing on the first relevant segment according to a first frame extraction frequency to obtain a first video frame sequence;

According to each frame-drawing frequency in the frame-drawing frequency sequence, the sub-segments in the second relevant segment are sequentially subjected to frame-drawing processing, and different sub-segments are extracted for each frame-drawing frequency during the frame-drawing process to obtain a second video frame sequence;

Wherein, each frame drawing frequency in the frame drawing frequency sequence is increasing or decreasing.

In one of the embodiments, the critical segment speed control module 408 is used to:

determining a target frame rate, the target frame rate being lower than the frame rate of the key segment;

converting the frame rate of the key segment to the target frame rate to obtain a slow-motion video segment.

In one of the embodiments, the fragment extraction module 406 is configured to:

Determining the critical time period during which the key event exists in the video;

Screening key segments in the video according to the key time period;

determining at least one relevant time period in the preceding preset time period of the critical time period, the subsequent preset time period of the critical time period, and the interval time period between each of the critical time periods;

Find relevant segments in the video according to the relevant time period.

In one of the embodiments, the event detection module 404 is configured to:

Obtain scene information of the video;

Detecting the action of the target object according to the scene information;

If it is determined that the action of the target object belongs to the preset action of the scene information, then determine the key event that the target object triggers the preset action.

In one of the embodiments, the event detection module 404 is configured to:

If the scene information represents a basketball scene, then determine that the preset action of the scene information is a shooting action;

According to the human body posture corresponding to the basketball scene information, extracting the posture key points of the target object from the video;

identifying whether the action of the target object belongs to a shooting action according to the key points of the posture;

The action of determining the target object belongs to the shooting action.

In one of the embodiments, the event detection module 404 is configured to:

determining a video frame of the video;

In the video picture, detect the area parameter of the target object; the area parameter includes the area of the target object in the video picture, or the ratio between the area of the target object and the area of the video picture;

When the area parameter is greater than a preset area parameter threshold, it is determined that the target object triggers a key event.

Each module in the above-mentioned video processing device may be fully or partially realized by software, hardware or a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, and can also be stored in the memory of the computer device in the form of software, so that the processor can invoke and execute the corresponding operations of the above-mentioned modules.

In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure may be as shown in FIG. 5 . The computer device includes a processor, a memory, an input/output interface, a communication interface, a display unit and an input device. Wherein, the processor, the memory and the input/output interface are connected through the system bus, and the communication interface, the display unit and the input device are connected to the system bus through the input/output interface. Wherein, the processor of the computer device is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input/output interface of the computer device is used for exchanging information between the processor and external devices. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be realized through WIFI, mobile cellular network, NFC (Near Field Communication) or other technologies. When the computer program is executed by a processor, a video processing method is realized. The display unit of the computer equipment is used to form a visually visible picture, and may be a display screen, a projection device or a virtual reality imaging device, the display screen may be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer equipment may be a display screen The touch layer covered on the upper surface may also be a button, a trackball or a touch pad arranged on the casing of the computer device, or an external keyboard, touch pad or mouse.

Those skilled in the art can understand that the structure shown in Figure 5 is only a block diagram of a part of the structure related to the solution of this application, and does not constitute a limitation to the computer equipment on which the solution of this application is applied. The specific computer equipment can be More or fewer components than shown in the figures may be included, or some components may be combined, or have a different arrangement of components.

In one embodiment, there is also provided a computer device, including a memory and a processor, where a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps in the foregoing method embodiments are implemented.

In one embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the steps in the foregoing method embodiments are implemented.

It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all It is information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with relevant laws, regulations and standards of relevant countries and regions.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above-mentioned embodiments can be completed by instructing related hardware through computer programs, and the computer programs can be stored in a non-volatile computer-readable memory In the medium, when the computer program is executed, it may include the processes of the embodiments of the above-mentioned methods. Wherein, any reference to storage, database or other media used in the various embodiments provided in the present application may include at least one of non-volatile and volatile storage. Non-volatile memory can include read-only memory (Read-Only Memory, ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive variable memory (ReRAM), magnetic variable memory (Magnetoresistive Random Access Memory, MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (Phase Change Memory, PCM), graphene memory, etc. The volatile memory may include random access memory (Random Access Memory, RAM) or external cache memory. As an illustration and not a limitation, RAM can be in various forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (Dynamic Random Access Memory, DRAM). The databases involved in the various embodiments provided in this application may include at least one of a relational database and a non-relational database. The non-relational database may include a blockchain-based distributed database, etc., but is not limited thereto. The processors involved in the various embodiments provided by this application can be general-purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, data processing logic devices based on quantum computing, etc., and are not limited to this.

The technical features of the above embodiments can be combined arbitrarily. To make the description concise, all possible combinations of the technical features in the above embodiments are not described. However, as long as there is no contradiction in the combination of these technical features, they should be It is considered to be within the range described in this specification.

The above-mentioned embodiments only express several implementation modes of the present application, and the description thereof is relatively specific and detailed, but should not be construed as limiting the patent scope of the present application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application should be determined by the appended claims.

Claims (14)

1. A method of video processing, the method comprising:

acquiring a video containing a target object;

detecting the target object triggering key event;

determining key fragments in the video according to the key events, and determining related fragments outside the key fragments from the video;

performing frame rate adjustment on the key segments to obtain slow motion video segments with target frame rates;

And synthesizing a target video based on the relevant segment and the slow motion video segment.

2. The method of claim 1, wherein the synthesizing the target video based on the relevant segments and the slow motion video segments comprises:

performing frame extraction processing on the relevant fragments to obtain a video frame sequence;

synthesizing the video frame sequence to obtain a video segment with the target frame rate;

and carrying out video synthesis on the video segment with the target frame rate and the slow motion video segment to obtain a target video.

3. The method of claim 2, wherein the relevant fragments comprise a first relevant fragment and a second phase Guan Pianduan;

the frame extraction processing is carried out on the relevant fragments to obtain a video frame sequence, which comprises the following steps:

performing frame extraction processing on the first relevant segment and the second relevant segment to obtain a first video frame sequence and a second video frame sequence;

the synthesizing the video frame sequence to obtain the video segment with the target frame rate comprises the following steps:

synthesizing the first video frame sequence and the second video frame sequence respectively to obtain a first video segment and a second video segment with target frame rates;

The video synthesizing the synthesized video clip and the slow motion video clip comprises:

and performing video synthesis on the first video clip, the slow motion video clip and the second video clip.

4. A method according to claim 3, wherein the performing the frame extraction on the first relevant segment and the second relevant segment to obtain a first video frame sequence and a second video frame sequence comprises:

performing frame extraction processing on the first relevant segment and the second relevant segment according to a first frame extraction frequency to obtain a first video frame sequence and a second video frame sequence; or,

performing frame extraction processing on the first relevant segment and the second relevant segment according to a second frame extraction frequency to obtain a first video frame sequence and a second video frame sequence;

wherein the first frame rate is greater than the second frame rate, and the first frame rate is determined based on a ratio between a frame rate of the key segment and the target frame rate.

5. A method according to claim 3, wherein the performing the frame extraction on the first relevant segment and the second relevant segment to obtain a first video frame sequence and a second video frame sequence comprises:

Performing frame extraction processing on the first related segment according to a first frame extraction frequency to obtain a first video frame sequence;

performing frame extraction processing on the second related segment according to a second frame extraction frequency to obtain a second video frame sequence;

wherein the first frame rate is greater than the second frame rate, and the first frame rate is determined based on a ratio between a frame rate of the key segment and the target frame rate.

6. A method according to claim 3, wherein the performing the frame extraction on the first relevant segment and the second relevant segment to obtain a first video frame sequence and a second video frame sequence comprises:

performing frame extraction processing on the first related segment according to a first frame extraction frequency to obtain a first video frame sequence;

sequentially carrying out frame extraction processing on the sub-segments in the second related segments according to each frame extraction frequency in the frame extraction frequency sequence, and extracting different sub-segments from each frame extraction frequency in the frame extraction process to obtain a second video frame sequence;

wherein each frame frequency in the frame frequency sequence is increased or decreased.

7. The method of claim 1, wherein performing frame rate adjustment on the key segments to obtain slow motion video segments at a target frame rate comprises:

Determining a target frame rate, the target frame rate being lower than the frame rate of the key segment;

and converting the frame rate of the key segment into the target frame rate to obtain the slow motion video segment.

8. The method of claim 1, wherein the determining key segments in a video from the key events and determining relevant segments outside the key segments from the video comprises:

determining a critical time period when the critical event exists in the video;

screening key fragments in the video according to the key time period;

determining at least one relevant time period in a previous preset time period of the key time period, a subsequent preset time period of the key time period and an interval time period between the key time periods;

and searching the relevant fragments in the video according to the relevant time period.

9. The method of claim 1, wherein the detecting the target object triggering a critical event comprises:

acquiring scene information of the video;

detecting the action of the target object according to the scene information;

and if the action of the target object is determined to belong to the preset action of the scene information, determining a key event of triggering the preset action by the target object.

10. The method of claim 9, wherein said detecting an action of said target object in accordance with said scene information comprises:

if the scene information characterizes a basketball scene, determining a preset action of the scene information as a shooting action;

extracting gesture key points of the target object from the video according to the human body gesture corresponding to the basketball scene information;

identifying whether the action of the target object belongs to a shooting action according to the gesture key points;

the determining that the action of the target object belongs to the preset action of the scene information comprises the following steps:

determining that the action of the target object belongs to the shooting action.

11. The method of claim 1, wherein the detecting the target object triggering a critical event comprises:

determining a video picture of the video;

detecting an area parameter of the target object in the video picture; the area parameter includes a target object area in the video picture or a ratio between the target object area and an area of the video picture;

and when the area parameter is larger than a preset area parameter threshold, determining that the target object triggers a key event.

12. A video processing apparatus, the apparatus comprising:

the video acquisition module is used for acquiring a video containing a target object;

the event detection module is used for detecting the target object triggering key event;

the segment extraction module is used for determining key segments in the video according to the key events and determining related segments outside the key segments from the video;

the key segment speed regulation module is used for carrying out frame rate regulation on the key segments to obtain slow motion video segments with target frame rate;

and the video synthesis module is used for synthesizing a target video based on the relevant segment and the slow motion video segment.

13. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any one of claims 1 to 11 when the computer program is executed.

14. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method of any of claims 1 to 11.

Images