CN110781836A - Human body recognition method and device, computer equipment and storage medium - Google Patents
Human body recognition method and device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN110781836A CN110781836A CN201911032168.XA CN201911032168A CN110781836A CN 110781836 A CN110781836 A CN 110781836A CN 201911032168 A CN201911032168 A CN 201911032168A CN 110781836 A CN110781836 A CN 110781836A
- Authority
- CN
- China
- Prior art keywords
- sample
- neural network
- labels
- convolutional neural
- image data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/20—Scenes; Scene-specific elements in augmented reality scenes
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to a human body identification method, a human body identification device, computer equipment and a storage medium, wherein the method comprises the steps of acquiring real-time image data shot by a camera on an unmanned aerial vehicle to obtain image data to be identified; inputting image data to be recognized into a target recognition model for recognition to obtain a recognition result; sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle performs corresponding operation on the identification result; the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence labels. The image data with the labels is used as a sample set, the deep convolutional neural network is trained to obtain the target recognition model, the target recognition model is applicable to the unmanned aerial vehicle, the real-time image data shot by a camera on the unmanned aerial vehicle is input into the target recognition model for target detection and recognition, and the recognition accuracy of the human body can be improved.
Description
Technical Field
The present invention relates to a target recognition method, and more particularly, to a human body recognition method, apparatus, computer device, and storage medium.
Background
In the past decades, object recognition has been a widespread field of computer vision. The excellent performance of deep learning techniques in detecting and classifying objects in images has led to their widespread adoption in various applications.
Detecting a human body in an image captured from a video stream is also one of the modes of target recognition, but detecting a human body is a challenging task because the human body has a variable appearance and can adopt various postures, and in addition, a video lens acquired by an unmanned aerial vehicle is greatly different from an image acquired on the ground; therefore, the use of the standard technology is not simple, because the unmanned aerial vehicle belongs to a prostrate human body, and the typical distance from the video camera to the human body is usually greater than the conditions of standard ground such as office-type environment and surveillance camera, so the current target detection algorithm suitable for the ground is not suitable for the unmanned aerial vehicle to perform target recognition, and the accuracy of human body recognition cannot be improved at a certain height and under different postures and different illumination conditions.
Therefore, it is necessary to design a new method, which is applicable to an unmanned aerial vehicle to perform target detection and identification, and improve the identification accuracy of a human body.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a human body identification method, a human body identification device, computer equipment and a storage medium.
In order to achieve the purpose, the invention adopts the following technical scheme: the human body recognition method comprises the following steps:
acquiring real-time image data shot by a camera on an unmanned aerial vehicle to obtain image data to be identified;
inputting image data to be recognized into a target recognition model for recognition to obtain a recognition result;
sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle performs corresponding operation on the identification result;
the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels.
The further technical scheme is as follows: the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels, and comprises the following steps:
acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set;
constructing a deep convolutional neural network and a loss function;
segmenting the sample set to obtain a sample grid unit set;
carrying out convolution training on the deep convolution neural network by using the sample grid unit set to obtain a training result;
calculating loss values of the training results, the class labels, the positioning labels and the confidence degree labels by using a loss function;
judging whether the loss value exceeds a threshold value;
if so, adjusting parameters of the deep convolutional neural network, and returning the sample grid unit set to carry out convolutional training on the deep convolutional neural network to obtain a training result;
and if not, taking the deep convolutional neural network as a target recognition model.
The further technical scheme is as follows: the convolutional training of the deep convolutional neural network by the sample grid unit set to obtain a training result, including:
predicting a boundary box of the sample network unit set by utilizing a deep convolutional neural network to obtain a plurality of sample boundary boxes;
carrying out prediction confidence score on the boundary boxes of the plurality of samples by utilizing a deep convolutional neural network to obtain a confidence value;
predicting relevant numerical values of the sample boundary boxes by using a deep convolutional neural network to obtain a plurality of positioning coordinates, the size of the sample boundary boxes and the probability of the human body appearing in the sample boundary boxes;
determining the probability of each category for the sample network unit set by using a logic classifier in the deep convolutional neural network to obtain the category probability;
obtaining the category corresponding to the sample network unit according to the category probability;
and integrating the confidence value, a plurality of positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
The further technical scheme is as follows: the calculating the loss value of the training result, the class label, the positioning label and the confidence label by using the loss function includes:
calculating corresponding variances for classes and class labels corresponding to the sample network units by using the loss function to obtain a classification loss value;
calculating corresponding variances for the positioning coordinates and the positioning labels by using a loss function to obtain a positioning loss value;
calculating corresponding variance by using the loss function to the confidence value and the confidence label to obtain a confidence loss value;
and integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
The further technical scheme is as follows: the deep convolutional neural network includes 13 convolutional layers.
The further technical scheme is as follows: the input size of the deep convolutional neural network is 416 × 416 image data.
The present invention also provides a human body recognition apparatus comprising:
the real-time image data acquisition unit is used for acquiring real-time image data shot by a camera on the unmanned aerial vehicle so as to obtain image data to be identified;
the data identification unit is used for inputting the image data to be identified into the target identification model for identification so as to obtain an identification result;
and the result sending unit is used for sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle carries out corresponding operation on the identification result.
The further technical scheme is as follows: further comprising:
and the model acquisition unit is used for training the deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels to obtain a target recognition model.
The invention also provides computer equipment which comprises a memory and a processor, wherein the memory is stored with a computer program, and the processor realizes the method when executing the computer program.
The invention also provides a storage medium storing a computer program which, when executed by a processor, is operable to carry out the method as described above.
Compared with the prior art, the invention has the beneficial effects that: the method comprises the steps of training a deep convolutional neural network by taking image data with labels as a sample set, and training the deep convolutional neural network from three aspects of positioning, classification and confidence coefficient to obtain a target recognition model, wherein the target recognition model is applicable to an unmanned aerial vehicle.
The invention is further described below with reference to the accompanying drawings and specific embodiments.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.
Fig. 1 is a schematic view of an application scenario of a human body recognition method according to an embodiment of the present invention;
fig. 2 is a schematic flow chart of a human body recognition method according to an embodiment of the present invention;
fig. 3 is a schematic sub-flow chart of a human body recognition method according to an embodiment of the present invention;
fig. 4 is a schematic sub-flow chart of a human body recognition method according to an embodiment of the present invention;
fig. 5 is a schematic sub-flow chart of a human body recognition method according to an embodiment of the present invention;
fig. 6 is a schematic block diagram of a human body recognition apparatus provided by an embodiment of the present invention;
FIG. 7 is a schematic block diagram of a computer device provided by an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
It will be understood that the terms "comprises" and/or "comprising," when used in this specification and the appended claims, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
It is also to be understood that the terminology used in the description of the invention herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in the specification of the present invention and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
It should be further understood that the term "and/or" as used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.
Referring to fig. 1 and fig. 2, fig. 1 is a schematic view of an application scenario of the human body recognition method according to the embodiment of the present invention. Fig. 2 is a schematic flow chart of a human body recognition method according to an embodiment of the present invention. The human body identification method is applied to a server. This server carries out the data interaction with unmanned aerial vehicle and camera, realizes carrying out human body detection and discernment after the camera on the unmanned aerial vehicle acquires real-time image data to obtain the discernment result, and send to unmanned aerial vehicle, in order to supply unmanned aerial vehicle as other operations of avoiding crowd such as input material.
Fig. 2 is a schematic flow chart of a human body recognition method according to an embodiment of the present invention. As shown in fig. 2, the method includes the following steps S110 to S150.
And S110, acquiring real-time image data shot by a camera on the unmanned aerial vehicle to obtain image data to be identified.
In this embodiment, the image data to be recognized refers to real-time image data of a certain specified area captured by a camera on the drone.
The human body on the ground is detected and identified from the angle of the unmanned aerial vehicle, and each frame of real-time video captured by the camera on the unmanned aerial vehicle can be simultaneously or sequentially input into the target identification model for identification.
And S120, inputting the image data to be recognized into the target recognition model for recognition to obtain a recognition result.
In this embodiment, the recognition result includes a category, a size of the grid cell, a number of bounding boxes, a confidence, and positioning information.
The target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels.
In an embodiment, referring to fig. 3, the step S120 may include steps S121 to S128.
S121, acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set.
In this embodiment, the sample set refers to a cocoa data set or a road crowd data set, which can be used as data for training the deep convolutional neural network.
Specifically, the sample set is generally divided into a training set and a testing set, wherein the training set is used for training the deep convolutional neural network, and the testing set is used for testing the recognition accuracy of the target recognition model, so as to obtain the target recognition model with high recognition accuracy.
And S122, constructing a deep convolutional neural network and a loss function.
In the present embodiment, the deep convolutional neural network includes 13 convolutional layers, and the input size of the deep convolutional neural network is 416 × 416 image data.
In particular, since the number of available data sets, i.e., sample sets, is limited, the transfer learning may be used for a pre-trained target recognition model to train a cable news web to detect road crowds, the training of the target recognition model is performed by initializing and fixing the depth of the pre-training weights of the last convolutional layer, the target recognition model achieves an average accuracy of about 67%, and the average is useful for evaluating model positioning performance, target monitoring model performance, and segmentation model performance.
And S123, segmenting the sample set to obtain a sample grid unit set.
In this embodiment, the sample grid cell set refers to a sample grid cell set obtained by performing image segmentation on a sample set to obtain N cells in a grid number of S × S. If the center of the body is located within a grid cell, then this particular cell is responsible for the detection of the body.
And S124, carrying out convolution training on the deep convolution neural network by the sample grid unit set to obtain a training result.
In this embodiment, the training result includes a confidence value, a plurality of positioning coordinates, a sample bounding box size, a probability of a human body appearing in the sample bounding box, and a category corresponding to the sample network unit.
In an embodiment, referring to fig. 4, the step S124 may include steps S1241 to S1246.
S1241, predicting the bounding box of the sample network unit set by using a deep convolutional neural network to obtain a plurality of sample bounding boxes.
In this embodiment, the sample bounding box refers to a box formed by the boundary positioning of each sample network unit.
S1242, carrying out prediction confidence score on the sample bounding boxes by using a deep convolutional neural network to obtain a confidence value.
In this embodiment, the confidence value is used to reflect the confidence of the target recognition model on the human body contained in the sample bounding box, and the accuracy of the predicted sample bounding box.
The formal definition of the confidence value is
Wherein Conf is a confidence value, expressed as a percentage fraction;
the intersection of the basic detection frames, namely the intersection of the sample network units, Pr is a prediction frame containing the object, namely a sample boundary frame containing the human body, and the acquisition is carried out according to the intersection of the prediction frame and the basic fact.
S1243, predicting relevant numerical values of the sample bounding boxes by utilizing a deep convolutional neural network to obtain a plurality of positioning coordinates, the sizes of the sample bounding boxes and the probability of the human body appearing in the sample bounding boxes.
Each sample bounding box contains 5 predictors, namely the coordinates of the (x, y) bounding box center, the box size, i.e. width and height, and the probability that a human body appears in the sample bounding box. All these predicted values are normalized values, i.e. between 0 and 1. The bounding box center coordinates (x, y) are calculated relative to the boundaries of the grid cells, while the width and height are predicted relative to the entire image data, so that the prediction of the confidence value represents a factor between the ground truth and the prediction box.
S1244, determining the probability of each category for the sample network unit set by using a logic classifier in the deep convolutional neural network to obtain the category probability.
In the present embodiment, the category probability refers to a human body category probability and a non-human body category probability.
The class probability is influenced by a sample network unit containing a human body, multi-label classification is used by replacing a softmax function with an independent logic classifier so as to calculate the probability that an input belongs to a specific label, and further the accuracy of the deep convolutional neural network for calculating the class probability of the sample network unit set is improved.
S1245, obtaining the category corresponding to the sample network unit according to the category probability.
In this embodiment, when the class probability is higher than a certain threshold, it indicates that the class corresponding to the sample network element is a human body, otherwise, it indicates that the class corresponding to the sample network element is a non-human body.
S1246, integrating the confidence value, the plurality of positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
During the training process, the loss function of the end-to-end model should deal with both classification and detection problems, which determines the learning and testing accuracy of the detector.
In the present embodiment, the deep convolutional neural network is a neural network adopting a mini YOLO V3 architecture. The miniature YOLO V3 architecture incorporates an improvement over YOLO, which differs from detection by predicting the human body at two different scales while extracting features from the underlying network. Compared with YOLO, the recognition accuracy rate is higher.
And S125, calculating loss values of the training result, the class label, the positioning label and the confidence label by using a loss function.
In this embodiment, the loss values include a classification loss value, a localization loss value, and a confidence loss value.
The calculation of the loss value is done by using the sum of squares error between the ground truth and the prediction.
In one embodiment, referring to fig. 5, the step S125 may include steps S1251 to S1254.
S1251, calculating corresponding variances of the classes and class labels corresponding to the sample network units by using the loss function to obtain a classification loss value.
In the present embodiment, the classification loss value refers to an error in prediction accuracy.
S1252, calculating corresponding variances of the plurality of positioning coordinates and the positioning labels by using the loss function to obtain a positioning loss value.
In this embodiment, the positioning loss value refers to an error between the positioning coordinates of the predicted sample bounding box and the ground true value, i.e., the positioning tag.
And S1253, calculating corresponding variances for the confidence values and the confidence labels by using the loss function to obtain confidence loss values.
In this embodiment, the confidence loss value refers to the degree of error of human appearance in the sample bounding box.
S1254, integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
And S126, judging whether the loss value exceeds a threshold value.
When any numerical value of the classification loss value, the positioning loss value and the confidence coefficient loss value exceeds a threshold value, the loss value is indicated to exceed the threshold value, and only if the classification loss value, the positioning loss value and the confidence coefficient loss value do not exceed the threshold value, the loss value is indicated not to exceed the threshold value.
And S127, if so, adjusting parameters of the deep convolutional neural network, and returning to the step S124.
The difference between the label and the actual numerical value is obtained according to the loss function, and then the parameter of the depth convolution neural network is adjusted according to the difference, so that the difference between the label and the actual numerical value meets the set threshold value, the human body can be accurately identified by the whole target identification model, the unmanned aerial vehicle is driven to search for unmanned areas to put in materials according to the identification result, or the positions where rescuers are needed to put in materials nearby, and the like.
And S128, if not, taking the deep convolutional neural network as a target recognition model.
The deep convolutional neural network encompasses the basic model as a feature extractor, the resulting tensor of the coding bounding box, the probability of the human body being present in the grid cell, and the class prediction.
S130, the identification result is sent to the unmanned aerial vehicle, so that the unmanned aerial vehicle carries out corresponding operation on the identification result.
The final output of the whole process of identifying the image data to be identified by the target identification model is the encoded sxs (B × 5+ C) tensor, wherein sxs is the size of the grid cell, B is the number of the bounding boxes, and C is the number of the network marker classes, and overall, the identification result includes data such as positioning, class and the like.
Deep learning techniques can extract very useful information from aerial images, and convolutional neural networks are the most advanced in advanced applications of deep learning for a wide range of tasks, such as image or video processing, speech recognition, text digitization, and the like.
According to the human body identification method, the image data with the labels are used as the sample set, the deep convolutional neural network is trained, and the deep convolutional neural network is trained from the aspects of positioning, classification and confidence coefficient to obtain the target identification model, the target identification model is applicable to the unmanned aerial vehicle, in the actual use process, the real-time image data shot by the camera on the unmanned aerial vehicle is input into the target identification model for target detection and identification, and the identification accuracy of the human body can be improved.
Fig. 6 is a schematic block diagram of a human body recognition apparatus 300 according to an embodiment of the present invention. As shown in fig. 6, the present invention also provides a human body recognition apparatus 300 corresponding to the above human body recognition method. The human body recognition apparatus 300 includes a unit for performing the above-described human body recognition method, and the apparatus may be configured in a server.
Specifically, referring to fig. 6, the human body recognition apparatus 300 includes:
a real-time image data acquiring unit 301, configured to acquire real-time image data captured by a camera on an unmanned aerial vehicle, so as to obtain image data to be identified;
the data identification unit 302 is used for inputting image data to be identified into a target identification model for identification so as to obtain an identification result;
a result sending unit 303, configured to send the identification result to the drone, so that the drone performs corresponding operation on the identification result
In one embodiment, the method further comprises:
and the model acquisition unit is used for training the deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels to obtain a target recognition model.
In one embodiment, the model obtaining unit includes:
the sample set acquisition subunit is used for acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set;
the building subunit is used for building a deep convolutional neural network and a loss function;
the dividing subunit is used for dividing the sample set to obtain a sample grid unit set;
the convolution training subunit is used for carrying out convolution training on the deep convolution neural network by the sample grid unit set so as to obtain a training result;
the loss value calculation operator unit is used for calculating a loss value of the training result, the class label, the positioning label and the confidence label by using a loss function;
a judging subunit, configured to judge whether the loss value exceeds a threshold value;
the parameter adjusting subunit is used for adjusting the parameters of the deep convolutional neural network if the parameters of the deep convolutional neural network are the same as the parameters of the sample grid unit set, and returning the sample grid unit set to carry out convolutional training on the deep convolutional neural network so as to obtain a training result;
and the model forming subunit is used for taking the deep convolutional neural network as a target identification model if the target identification model is not the deep convolutional neural network.
In one embodiment, the convolution training subunit includes:
the boundary box acquisition module is used for predicting the boundary box of the sample network unit set by utilizing the deep convolutional neural network so as to obtain a plurality of sample boundary boxes;
the score prediction module is used for carrying out prediction confidence score on the sample bounding boxes by utilizing a deep convolutional neural network so as to obtain a confidence value;
the numerical value prediction module is used for predicting relevant numerical values of the sample boundary boxes by utilizing the deep convolutional neural network so as to obtain a plurality of positioning coordinates, the sizes of the sample boundary boxes and the probability of the human body appearing in the sample boundary boxes;
the class probability determining module is used for determining the probability of each class for the sample network unit set by utilizing a logic classifier in the deep convolutional neural network so as to obtain the class probability;
the category acquisition module is used for acquiring categories corresponding to the sample network units according to the category probability;
and the first integration module is used for integrating the confidence value, the positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
In an embodiment, the loss value operator unit comprises:
the classification loss value acquisition module is used for calculating corresponding variances for classes and class labels corresponding to the sample network units by using the loss function so as to obtain a classification loss value;
the positioning loss value acquisition module is used for calculating corresponding variances of the positioning coordinates and the positioning labels by using a loss function so as to obtain a positioning loss value;
the confidence coefficient loss value acquisition module is used for calculating corresponding variances for the confidence coefficient values and the confidence coefficient labels by using the loss function so as to obtain confidence coefficient loss values;
and the second integration module is used for integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
It should be noted that, as can be clearly understood by those skilled in the art, the specific implementation processes of the human body recognition apparatus 300 and each unit may refer to the corresponding descriptions in the foregoing method embodiments, and for convenience and brevity of description, no further description is provided herein.
The above-described human body recognition apparatus 300 may be implemented in the form of a computer program that can be run on a computer device as shown in fig. 7.
Referring to fig. 7, fig. 7 is a schematic block diagram of a computer device according to an embodiment of the present application. The computer device 500 is a server.
Referring to fig. 7, the computer device 500 includes a processor 502, memory, and a network interface 505 connected by a system bus 501, where the memory may include a non-volatile storage medium 503 and an internal memory 504.
The non-volatile storage medium 503 may store an operating system 5031 and a computer program 5032. The computer program 5032 comprises program instructions that, when executed, cause the processor 502 to perform a human recognition method.
The processor 502 is used to provide computing and control capabilities to support the operation of the overall computer device 500.
The internal memory 504 provides an environment for running the computer program 5032 in the non-volatile storage medium 503, and when the computer program 5032 is executed by the processor 502, the processor 502 can be caused to execute a human body recognition method.
The network interface 505 is used for network communication with other devices. Those skilled in the art will appreciate that the configuration shown in fig. 7 is a block diagram of only a portion of the configuration associated with the present application and does not constitute a limitation of the computer device 500 to which the present application may be applied, and that a particular computer device 500 may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.
Wherein the processor 502 is configured to run the computer program 5032 stored in the memory to implement the following steps:
acquiring real-time image data shot by a camera on an unmanned aerial vehicle to obtain image data to be identified;
inputting image data to be recognized into a target recognition model for recognition to obtain a recognition result;
sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle performs corresponding operation on the identification result;
the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels.
In an embodiment, when the processor 502 implements the step of training the deep convolutional neural network by using a plurality of image data with class labels, positioning labels, and confidence labels, the following steps are implemented:
acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set;
constructing a deep convolutional neural network and a loss function;
segmenting the sample set to obtain a sample grid unit set;
carrying out convolution training on the deep convolution neural network by using the sample grid unit set to obtain a training result;
calculating loss values of the training results, the class labels, the positioning labels and the confidence degree labels by using a loss function;
judging whether the loss value exceeds a threshold value;
if so, adjusting parameters of the deep convolutional neural network, and returning the sample grid unit set to carry out convolutional training on the deep convolutional neural network to obtain a training result;
and if not, taking the deep convolutional neural network as a target recognition model.
Wherein the deep convolutional neural network comprises 13 convolutional layers.
The input size of the deep convolutional neural network is 416 × 416 image data.
In an embodiment, when the step of performing convolution training on the deep convolutional neural network by using the sample grid unit set to obtain a training result is implemented by the processor 502, the following steps are specifically implemented:
predicting a boundary box of the sample network unit set by utilizing a deep convolutional neural network to obtain a plurality of sample boundary boxes;
carrying out prediction confidence score on the boundary boxes of the plurality of samples by utilizing a deep convolutional neural network to obtain a confidence value;
predicting relevant numerical values of the sample boundary boxes by using a deep convolutional neural network to obtain a plurality of positioning coordinates, the size of the sample boundary boxes and the probability of the human body appearing in the sample boundary boxes;
determining the probability of each category for the sample network unit set by using a logic classifier in the deep convolutional neural network to obtain the category probability;
obtaining the category corresponding to the sample network unit according to the category probability;
and integrating the confidence value, a plurality of positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
In an embodiment, when the step of calculating the loss value of the training result, the class label, the positioning label, and the confidence label by using the loss function is implemented, the processor 502 specifically implements the following steps:
calculating corresponding variances for classes and class labels corresponding to the sample network units by using the loss function to obtain a classification loss value;
calculating corresponding variances for the positioning coordinates and the positioning labels by using a loss function to obtain a positioning loss value;
calculating corresponding variance by using the loss function to the confidence value and the confidence label to obtain a confidence loss value;
and integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
It should be understood that, in the embodiment of the present Application, the Processor 502 may be a Central Processing Unit (CPU), and the Processor 502 may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, and the like. Wherein a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
It will be understood by those skilled in the art that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program instructing associated hardware. The computer program includes program instructions, and the computer program may be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the flow steps of the embodiments of the method described above.
Accordingly, the present invention also provides a storage medium. The storage medium may be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program, when executed by a processor, causes the processor to perform the steps of:
acquiring real-time image data shot by a camera on an unmanned aerial vehicle to obtain image data to be identified;
inputting image data to be recognized into a target recognition model for recognition to obtain a recognition result;
sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle performs corresponding operation on the identification result;
the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels.
In an embodiment, when the processor executes the computer program to implement the step of training the deep convolutional neural network by using a plurality of image data with class labels, positioning labels and confidence labels, the processor specifically implements the following steps:
acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set;
constructing a deep convolutional neural network and a loss function;
segmenting the sample set to obtain a sample grid unit set;
carrying out convolution training on the deep convolution neural network by using the sample grid unit set to obtain a training result;
calculating loss values of the training results, the class labels, the positioning labels and the confidence degree labels by using a loss function;
judging whether the loss value exceeds a threshold value;
if so, adjusting parameters of the deep convolutional neural network, and returning the sample grid unit set to carry out convolutional training on the deep convolutional neural network to obtain a training result;
and if not, taking the deep convolutional neural network as a target recognition model.
Wherein the deep convolutional neural network comprises 13 convolutional layers.
The input size of the deep convolutional neural network is 416 × 416 image data.
In an embodiment, when the processor executes the computer program to perform the step of performing convolutional training on the deep convolutional neural network by using the sample grid cell set to obtain a training result, the following steps are specifically implemented:
predicting a boundary box of the sample network unit set by utilizing a deep convolutional neural network to obtain a plurality of sample boundary boxes;
carrying out prediction confidence score on the boundary boxes of the plurality of samples by utilizing a deep convolutional neural network to obtain a confidence value;
predicting relevant numerical values of the sample boundary boxes by using a deep convolutional neural network to obtain a plurality of positioning coordinates, the size of the sample boundary boxes and the probability of the human body appearing in the sample boundary boxes;
determining the probability of each category for the sample network unit set by using a logic classifier in the deep convolutional neural network to obtain the category probability;
obtaining the category corresponding to the sample network unit according to the category probability;
and integrating the confidence value, a plurality of positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
In an embodiment, when the processor executes the computer program to implement the step of calculating the loss value using the loss function for the training result, the category label, the positioning label, and the confidence label, the following steps are specifically implemented:
calculating corresponding variances for classes and class labels corresponding to the sample network units by using the loss function to obtain a classification loss value;
calculating corresponding variances for the positioning coordinates and the positioning labels by using a loss function to obtain a positioning loss value;
calculating corresponding variance by using the loss function to the confidence value and the confidence label to obtain a confidence loss value;
and integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
The storage medium may be a usb disk, a removable hard disk, a Read-Only Memory (ROM), a magnetic disk, or an optical disk, which can store various computer readable storage media.
Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the examples described in connection with the embodiments disclosed herein may be embodied in electronic hardware, computer software, or combinations of both, and that the components and steps of the examples have been described in a functional general in the foregoing description for the purpose of illustrating clearly the interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
In the embodiments provided in the present invention, it should be understood that the disclosed apparatus and method may be implemented in other ways. For example, the above-described apparatus embodiments are merely illustrative. For example, the division of each unit is only one logic function division, and there may be another division manner in actual implementation. For example, various elements or components may be combined or may be integrated into another system, or some features may be omitted, or not implemented.
The steps in the method of the embodiment of the invention can be sequentially adjusted, combined and deleted according to actual needs. The units in the device of the embodiment of the invention can be merged, divided and deleted according to actual needs. In addition, functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may exist alone physically, or two or more units are integrated into one unit.
The integrated unit, if implemented in the form of a software functional unit and sold or used as a stand-alone product, may be stored in a storage medium. Based on such understanding, the technical solution of the present invention essentially or partially contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes instructions for causing a computer device (which may be a personal computer, a terminal, or a network device) to execute all or part of the steps of the method according to the embodiments of the present invention.
While the invention has been described with reference to specific embodiments, the invention is not limited thereto, and various equivalent modifications and substitutions can be easily made by those skilled in the art within the technical scope of the invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.
Claims (10)
1. The human body recognition method is characterized by comprising the following steps:
acquiring real-time image data shot by a camera on an unmanned aerial vehicle to obtain image data to be identified;
inputting image data to be recognized into a target recognition model for recognition to obtain a recognition result;
sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle performs corresponding operation on the identification result;
the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels.
2. The human body recognition method of claim 1, wherein the target recognition model is obtained by training a deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence labels, and comprises the following steps:
acquiring a plurality of image data with category labels, positioning labels and confidence labels to obtain a sample set;
constructing a deep convolutional neural network and a loss function;
segmenting the sample set to obtain a sample grid unit set;
carrying out convolution training on the deep convolution neural network by using the sample grid unit set to obtain a training result;
calculating loss values of the training results, the class labels, the positioning labels and the confidence degree labels by using a loss function;
judging whether the loss value exceeds a threshold value;
if so, adjusting parameters of the deep convolutional neural network, and returning the sample grid unit set to carry out convolutional training on the deep convolutional neural network to obtain a training result;
and if not, taking the deep convolutional neural network as a target recognition model.
3. The human body recognition method of claim 2, wherein the convolutional training of the deep convolutional neural network with the sample grid cell set to obtain a training result comprises:
predicting a boundary box of the sample network unit set by utilizing a deep convolutional neural network to obtain a plurality of sample boundary boxes;
carrying out prediction confidence score on the boundary boxes of the plurality of samples by utilizing a deep convolutional neural network to obtain a confidence value;
predicting relevant numerical values of the sample boundary boxes by using a deep convolutional neural network to obtain a plurality of positioning coordinates, the size of the sample boundary boxes and the probability of the human body appearing in the sample boundary boxes;
determining the probability of each category for the sample network unit set by using a logic classifier in the deep convolutional neural network to obtain the category probability;
obtaining the category corresponding to the sample network unit according to the category probability;
and integrating the confidence value, a plurality of positioning coordinates, the size of the sample boundary box, the probability of the human body appearing in the sample boundary box and the corresponding category of the sample network unit to obtain a training result.
4. The human body recognition method according to claim 3, wherein the calculating the loss value of the training result, the class label, the location label and the confidence label by using the loss function comprises:
calculating corresponding variances for classes and class labels corresponding to the sample network units by using the loss function to obtain a classification loss value;
calculating corresponding variances for the positioning coordinates and the positioning labels by using a loss function to obtain a positioning loss value;
calculating corresponding variance by using the loss function to the confidence value and the confidence label to obtain a confidence loss value;
and integrating the classification loss value, the positioning loss value and the confidence coefficient loss value to obtain a loss value.
5. The human recognition method of claim 2, wherein the deep convolutional neural network comprises 13 convolutional layers.
6. The human recognition method of claim 2, wherein an input size of the deep convolutional neural network is 416 x 416 image data.
7. A human body recognition device, comprising:
the real-time image data acquisition unit is used for acquiring real-time image data shot by a camera on the unmanned aerial vehicle so as to obtain image data to be identified;
the data identification unit is used for inputting the image data to be identified into the target identification model for identification so as to obtain an identification result;
and the result sending unit is used for sending the identification result to the unmanned aerial vehicle so that the unmanned aerial vehicle carries out corresponding operation on the identification result.
8. The human recognition device according to claim 7, further comprising:
and the model acquisition unit is used for training the deep convolutional neural network through a plurality of image data with class labels, positioning labels and confidence degree labels to obtain a target recognition model.
9. A computer device, characterized in that the computer device comprises a memory, on which a computer program is stored, and a processor, which when executing the computer program implements the method according to any of claims 1 to 6.
10. A storage medium, characterized in that the storage medium stores a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911032168.XA CN110781836A (en) | 2019-10-28 | 2019-10-28 | Human body recognition method and device, computer equipment and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201911032168.XA CN110781836A (en) | 2019-10-28 | 2019-10-28 | Human body recognition method and device, computer equipment and storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110781836A true CN110781836A (en) | 2020-02-11 |
Family
ID=69387173
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201911032168.XA Pending CN110781836A (en) | 2019-10-28 | 2019-10-28 | Human body recognition method and device, computer equipment and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110781836A (en) |
Cited By (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111401241A (en) * | 2020-03-16 | 2020-07-10 | 中国民用航空飞行学院 | Search method and device for maritime civil aircraft based on convolutional neural network |
| CN111461209A (en) * | 2020-03-30 | 2020-07-28 | 深圳市凯立德科技股份有限公司 | Model training device and method |
| CN111598942A (en) * | 2020-03-12 | 2020-08-28 | 中国电力科学研究院有限公司 | A method and system for automatic positioning of power facility meters |
| CN111666890A (en) * | 2020-06-08 | 2020-09-15 | 平安科技（深圳）有限公司 | Spine deformation crowd identification method and device, computer equipment and storage medium |
| CN111695592A (en) * | 2020-04-27 | 2020-09-22 | 平安科技（深圳）有限公司 | Image identification method and device based on deformable convolution and computer equipment |
| CN112001498A (en) * | 2020-08-14 | 2020-11-27 | 苏州浪潮智能科技有限公司 | Data identification method and device based on quantum computer and readable storage medium |
| CN112016475A (en) * | 2020-08-31 | 2020-12-01 | 支付宝(杭州)信息技术有限公司 | Human body detection and identification method and device |
| CN112036331A (en) * | 2020-09-03 | 2020-12-04 | 腾讯科技（深圳）有限公司 | Training method, device and equipment of living body detection model and storage medium |
| CN112184756A (en) * | 2020-09-30 | 2021-01-05 | 北京理工大学 | Single-target rapid detection method based on deep learning |
| CN112257579A (en) * | 2020-10-21 | 2021-01-22 | 平安科技（深圳）有限公司 | Model training method, action recognition method, device, equipment and storage medium |
| CN112349150A (en) * | 2020-11-19 | 2021-02-09 | 飞友科技有限公司 | Video acquisition method and system for airport flight guarantee time node |
| CN112434715A (en) * | 2020-12-10 | 2021-03-02 | 腾讯科技（深圳）有限公司 | Target identification method and device based on artificial intelligence and storage medium |
| CN112668509A (en) * | 2020-12-31 | 2021-04-16 | 深圳云天励飞技术股份有限公司 | Training method and recognition method of social relationship recognition model and related equipment |
| CN112966618A (en) * | 2021-03-11 | 2021-06-15 | 京东数科海益信息科技有限公司 | Dressing identification method, device, equipment and computer readable medium |
| CN113111729A (en) * | 2021-03-23 | 2021-07-13 | 广州大学 | Training method, recognition method, system, device and medium of personnel recognition model |
| CN114359670A (en) * | 2021-12-28 | 2022-04-15 | 深圳中智永浩机器人有限公司 | Unstructured data labeling method and device, computer equipment and storage medium |
| CN114359967A (en) * | 2022-01-10 | 2022-04-15 | 杭州巨岩欣成科技有限公司 | Method and device for detecting drowning-preventing human body target in swimming pool, computer equipment and storage medium |
| CN114627501A (en) * | 2022-03-11 | 2022-06-14 | 深圳中智永浩机器人有限公司 | Moving target detection method and device, computer equipment and storage medium |
| CN114663717A (en) * | 2020-12-22 | 2022-06-24 | 顺丰科技有限公司 | Packaging material visual identification method, device, computer equipment and storage medium |
| CN116205969A (en) * | 2023-05-06 | 2023-06-02 | 深圳赛桥生物创新技术有限公司 | Sample liquid volume identification method, system and medium |
| CN117975144A (en) * | 2024-02-02 | 2024-05-03 | 北京视觉世界科技有限公司 | Target information identification method, device, equipment and storage medium |
| CN119341841A (en) * | 2024-12-19 | 2025-01-21 | 北京简网科技有限公司 | User behavior identification method, system, electronic device and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108921152A (en) * | 2018-06-29 | 2018-11-30 | 清华大学 | English character cutting method and device based on object detection network |
| CN109447033A (en) * | 2018-11-14 | 2019-03-08 | 北京信息科技大学 | Vehicle front obstacle detection method based on YOLO |
| CN109711437A (en) * | 2018-12-06 | 2019-05-03 | 武汉三江中电科技有限责任公司 | A kind of transformer part recognition methods based on YOLO network model |
| CN109815886A (en) * | 2019-01-21 | 2019-05-28 | 南京邮电大学 | A pedestrian and vehicle detection method and system based on improved YOLOv3 |
| US20190311203A1 (en) * | 2018-04-09 | 2019-10-10 | Accenture Global Solutions Limited | Aerial monitoring system and method for identifying and locating object features |
| CN110321775A (en) * | 2019-04-08 | 2019-10-11 | 武汉理工大学 | A kind of drowning man's autonomous classification method waterborne based on multi-rotor unmanned aerial vehicle |
| CN110348304A (en) * | 2019-06-06 | 2019-10-18 | 武汉理工大学 | A kind of maritime affairs distress personnel search system being equipped on unmanned plane and target identification method |
-
2019
- 2019-10-28 CN CN201911032168.XA patent/CN110781836A/en active Pending
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190311203A1 (en) * | 2018-04-09 | 2019-10-10 | Accenture Global Solutions Limited | Aerial monitoring system and method for identifying and locating object features |
| CN108921152A (en) * | 2018-06-29 | 2018-11-30 | 清华大学 | English character cutting method and device based on object detection network |
| CN109447033A (en) * | 2018-11-14 | 2019-03-08 | 北京信息科技大学 | Vehicle front obstacle detection method based on YOLO |
| CN109711437A (en) * | 2018-12-06 | 2019-05-03 | 武汉三江中电科技有限责任公司 | A kind of transformer part recognition methods based on YOLO network model |
| CN109815886A (en) * | 2019-01-21 | 2019-05-28 | 南京邮电大学 | A pedestrian and vehicle detection method and system based on improved YOLOv3 |
| CN110321775A (en) * | 2019-04-08 | 2019-10-11 | 武汉理工大学 | A kind of drowning man's autonomous classification method waterborne based on multi-rotor unmanned aerial vehicle |
| CN110348304A (en) * | 2019-06-06 | 2019-10-18 | 武汉理工大学 | A kind of maritime affairs distress personnel search system being equipped on unmanned plane and target identification method |
Cited By (36)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598942A (en) * | 2020-03-12 | 2020-08-28 | 中国电力科学研究院有限公司 | A method and system for automatic positioning of power facility meters |
| CN111401241A (en) * | 2020-03-16 | 2020-07-10 | 中国民用航空飞行学院 | Search method and device for maritime civil aircraft based on convolutional neural network |
| CN111461209A (en) * | 2020-03-30 | 2020-07-28 | 深圳市凯立德科技股份有限公司 | Model training device and method |
| CN111461209B (en) * | 2020-03-30 | 2024-04-09 | 深圳市凯立德科技股份有限公司 | Model training device and method |
| CN111695592B (en) * | 2020-04-27 | 2024-07-09 | 平安科技（深圳）有限公司 | Image recognition method and device based on deformable convolution and computer equipment |
| CN111695592A (en) * | 2020-04-27 | 2020-09-22 | 平安科技（深圳）有限公司 | Image identification method and device based on deformable convolution and computer equipment |
| CN111666890A (en) * | 2020-06-08 | 2020-09-15 | 平安科技（深圳）有限公司 | Spine deformation crowd identification method and device, computer equipment and storage medium |
| CN111666890B (en) * | 2020-06-08 | 2023-06-30 | 平安科技（深圳）有限公司 | Spine deformation crowd identification method and device, computer equipment and storage medium |
| CN112001498A (en) * | 2020-08-14 | 2020-11-27 | 苏州浪潮智能科技有限公司 | Data identification method and device based on quantum computer and readable storage medium |
| CN112001498B (en) * | 2020-08-14 | 2022-12-09 | 苏州浪潮智能科技有限公司 | Data identification method and device based on quantum computer and readable storage medium |
| CN112016475A (en) * | 2020-08-31 | 2020-12-01 | 支付宝(杭州)信息技术有限公司 | Human body detection and identification method and device |
| CN112016475B (en) * | 2020-08-31 | 2022-07-08 | 支付宝(杭州)信息技术有限公司 | Human body detection and identification method and device |
| CN112036331B (en) * | 2020-09-03 | 2024-04-09 | 腾讯科技（深圳）有限公司 | Living body detection model training method, device, equipment and storage medium |
| CN112036331A (en) * | 2020-09-03 | 2020-12-04 | 腾讯科技（深圳）有限公司 | Training method, device and equipment of living body detection model and storage medium |
| CN112184756A (en) * | 2020-09-30 | 2021-01-05 | 北京理工大学 | Single-target rapid detection method based on deep learning |
| CN112257579A (en) * | 2020-10-21 | 2021-01-22 | 平安科技（深圳）有限公司 | Model training method, action recognition method, device, equipment and storage medium |
| CN112257579B (en) * | 2020-10-21 | 2024-10-15 | 平安科技（深圳）有限公司 | Model training method, motion recognition method, device, equipment and storage medium |
| CN112349150A (en) * | 2020-11-19 | 2021-02-09 | 飞友科技有限公司 | Video acquisition method and system for airport flight guarantee time node |
| CN112434715A (en) * | 2020-12-10 | 2021-03-02 | 腾讯科技（深圳）有限公司 | Target identification method and device based on artificial intelligence and storage medium |
| CN112434715B (en) * | 2020-12-10 | 2022-07-22 | 腾讯科技（深圳）有限公司 | Target identification method and device based on artificial intelligence and storage medium |
| CN114663717A (en) * | 2020-12-22 | 2022-06-24 | 顺丰科技有限公司 | Packaging material visual identification method, device, computer equipment and storage medium |
| CN114663717B (en) * | 2020-12-22 | 2024-11-26 | 顺丰科技有限公司 | Packaging material visual identification method, device, computer equipment and storage medium |
| CN112668509B (en) * | 2020-12-31 | 2024-04-02 | 深圳云天励飞技术股份有限公司 | Training method and recognition method of social relation recognition model and related equipment |
| CN112668509A (en) * | 2020-12-31 | 2021-04-16 | 深圳云天励飞技术股份有限公司 | Training method and recognition method of social relationship recognition model and related equipment |
| CN112966618A (en) * | 2021-03-11 | 2021-06-15 | 京东数科海益信息科技有限公司 | Dressing identification method, device, equipment and computer readable medium |
| CN112966618B (en) * | 2021-03-11 | 2024-02-09 | 京东科技信息技术有限公司 | Dressing recognition method, apparatus, device and computer readable medium |
| CN113111729B (en) * | 2021-03-23 | 2023-08-18 | 广州大学 | Training method, recognition method, system, device and medium for personnel recognition model |
| CN113111729A (en) * | 2021-03-23 | 2021-07-13 | 广州大学 | Training method, recognition method, system, device and medium of personnel recognition model |
| CN114359670A (en) * | 2021-12-28 | 2022-04-15 | 深圳中智永浩机器人有限公司 | Unstructured data labeling method and device, computer equipment and storage medium |
| CN114359967A (en) * | 2022-01-10 | 2022-04-15 | 杭州巨岩欣成科技有限公司 | Method and device for detecting drowning-preventing human body target in swimming pool, computer equipment and storage medium |
| CN114627501A (en) * | 2022-03-11 | 2022-06-14 | 深圳中智永浩机器人有限公司 | Moving target detection method and device, computer equipment and storage medium |
| CN116205969B (en) * | 2023-05-06 | 2023-07-21 | 深圳赛桥生物创新技术有限公司 | Sample liquid volume identification method, system and medium |
| CN116205969A (en) * | 2023-05-06 | 2023-06-02 | 深圳赛桥生物创新技术有限公司 | Sample liquid volume identification method, system and medium |
| CN117975144A (en) * | 2024-02-02 | 2024-05-03 | 北京视觉世界科技有限公司 | Target information identification method, device, equipment and storage medium |
| CN119341841A (en) * | 2024-12-19 | 2025-01-21 | 北京简网科技有限公司 | User behavior identification method, system, electronic device and storage medium |
| CN119341841B (en) * | 2024-12-19 | 2025-04-01 | 北京简网科技有限公司 | User behavior recognition method, system, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110781836A (en) | Human body recognition method and device, computer equipment and storage medium | |
| CN106960195B (en) | Crowd counting method and device based on deep learning | |
| US11144786B2 (en) | Information processing apparatus, method for controlling information processing apparatus, and storage medium | |
| CN112132119B (en) | Passenger flow statistical method and device, electronic equipment and storage medium | |
| CN108229267B (en) | Object attribute detection, neural network training and region detection method and device | |
| WO2021051601A1 (en) | Method and system for selecting detection box using mask r-cnn, and electronic device and storage medium | |
| CN104615986B (en) | The method that pedestrian detection is carried out to the video image of scene changes using multi-detector | |
| CN111274926B (en) | Image data screening method, device, computer equipment and storage medium | |
| CN111860439A (en) | Unmanned aerial vehicle inspection image defect detection method, system and equipment | |
| KR20170026222A (en) | Method and device for classifying an object of an image and corresponding computer program product and computer-readable medium | |
| Xing et al. | Traffic sign recognition using guided image filtering | |
| CN114299457A (en) | Accumulated water depth detection method and device | |
| CN113409252A (en) | Obstacle detection method for overhead transmission line inspection robot | |
| CN113780145A (en) | Sperm morphology detection method, sperm morphology detection device, computer equipment and storage medium | |
| Yusuf et al. | Enhancing vehicle detection and tracking in UAV imagery: a pixel Labeling and particle filter approach | |
| CN114821441B (en) | A deep learning-based method for identifying moving targets at airports combined with ADS-B information | |
| KR101690050B1 (en) | Intelligent video security system | |
| CN115690554A (en) | Target identification method, system, electronic device and storage medium | |
| CN113610770B (en) | License plate recognition method, device and equipment | |
| Damodaran et al. | Overhead power line detection from aerial images using segmentation approaches | |
| CN117523428B (en) | Ground target detection method and device based on aircraft platform | |
| CN117037049B (en) | Image content detection method and system based on YOLOv5 deep learning | |
| CN108985216B (en) | Pedestrian head detection method based on multivariate logistic regression feature fusion | |
| Kaimkhani et al. | UAV with Vision to Recognise Vehicle Number Plates | |
| Aydin et al. | A new object detection and classification method for quality control based on segmentation and geometric features |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20200211 |
|
| RJ01 | Rejection of invention patent application after publication |