CN104281682A - File Classification System and Method - Google Patents

Abstract

A document classification system and method are disclosed, wherein the system includes a storage device, a receiving device, and a processor. The storage device stores at least one identification audio. The receiving device obtains an audio file or an image file. The processor compares a relevant audio frequency related to the sound file or the image file with the at least one identification audio frequency to generate a processing result, and then automatically classifies the sound file or the image file according to the processing result so as to eliminate the trouble of manually classifying the files.

Description

File Classification System and Method

technical field

The present invention relates to a classification technology, and in particular to a file classification system and a file classification method.

Background technique

The development of audio or video technology has been going on for some time, and most of them are researched and improved for sound or image capture technology, and seldom focus on the classification and storage of recorded files.

Generally speaking, the sound files or video files generated after recording or video recording are often stored in the same location, and the naming rules often use similar English and/or number combinations as file names; unless users rename themselves, Otherwise, it is difficult to determine the file content based on the file name alone. After a long period of time, when the number of files is huge and the user does not arrange them regularly, it is not easy to find a specific file from among the many files.

Contents of the invention

In order to solve the problem that files cannot be properly classified during recording or video recording, an aspect of the present invention provides a file classification system, which includes a storage device, a receiving device and a processor. The storage device stores at least one identification audio, and the receiving device obtains an audio file or an image file. The processor compares a relevant audio associated with the sound file or the image file with the at least one identified audio to generate a processing result, and then automatically classifies the sound file or the image file according to the processing result.

Another aspect of the present invention is to provide a file classification method, which includes the following steps: (A) store at least one identified audio; (B) obtain a sound file or an image file; (C) associate the sound file or A related audio of the image file is compared with the at least one identified audio to generate a processing result, and then automatically classify the audio file or the image file according to the processing result.

To sum up, the present invention is based on improving the file classification method, and provides a voice file recognition mechanism to quickly classify files without cumbersome operations.

The above-mentioned description will be described in detail in the following embodiments, and a further explanation will be provided for the technical solution of the present invention.

Description of drawings

In order to make the above and other objects, features, advantages and embodiments of the present invention more comprehensible, the accompanying drawings are described as follows:

FIG. 1 is a schematic diagram of a document classification system according to a first embodiment of the present invention;

Fig. 2 is a schematic diagram of a file classification system according to a second embodiment of the present invention;

Fig. 3 is a schematic diagram of a file classification system according to a third embodiment of the present invention;

FIG. 4 is a schematic diagram of a file classification system according to a fourth embodiment of the present invention; and

FIG. 5 is a flowchart of a file classification method according to a fifth embodiment of the present invention.

Detailed ways

In order to make the description of the present invention more detailed and complete, reference may be made to the accompanying drawings and various embodiments described below. However, the provided examples are not intended to limit the scope of the present invention; the description of the steps is not intended to limit the order in which they are executed, and any device that has the same effect produced by recombination is covered by the present invention scope.

Please refer to FIG. 1 , which illustrates a file classification system 100 according to a first embodiment of the present invention. The document classification system 100 includes a storage device 130 , a receiving device 140 and a processor 120 . In practice, the storage device 130 can be a hard disk, flash memory or other storage media, and the receiving device 140 can be at least one transmission port, which can be a wired and/or wireless transmission port according to actual needs, and can be based on actual needs The digital and/or analog transmission ports (such as: HDMI, USB, 3.5mm, etc.) are built-in or externally connected to recording and/or video recording devices, and the processor 120 can be a central processing unit, a microcontroller or other circuits.

Before recording or video recording, the user (such as: a speaker) can store a pre-recorded voice as the recognition audio 132 in the storage device 130 , and the storage device 130 stores the recognition audio 132 . After the recording or video recording is completed, the receiving device 140 receives the sound file or image file 110, the file classification system 100 can prompt the request or after the speaker selects the voice classification, the processor 120 will be associated with a sound file or the image file 110 The audio is compared with the recognized audio 132 to generate a processing result 122 , wherein the processing result 122 has suggested classification information, and then, the processor 120 automatically classifies the audio file or the video file 110 according to the processing result 122 . For example, if the relevant audio matches the identification audio 132, it means that the audio file or the video file 110 is the audio or video data of the lecturer himself. Therefore, the processor 120 automatically classifies the audio file or the video file 110 into The category customized by the lecturer is stored in the storage device 130, thereby avoiding complicated manual operations.

In order to further explain the recording method of the identification audio 132 of the present invention, please refer to FIG. 2 , which shows a file classification system 200 according to the second embodiment of the present invention. The file classification system 200 in FIG. 2 is in addition to adding an audio recording device 250 Other than that, the rest of the hardware is substantially the same as the document classification system 100 in FIG. 1 . In practice, the audio recording device 250 can be a microphone or other sound receiving devices. In addition, the audio recording device 250 can also be integrated with the built-in or external recording device of the receiving device 140 as the same device according to actual needs.

Before the audio file or image file 110 is recorded, the user (such as a speaker) can pre-record the above-mentioned recognition audio 132 through the built-in audio recording device 250 of the file classification system 200 as a reference sample for speech recognition, which is simple and convenient . After the recording or video recording is completed, the processor 120 compares the relevant audio associated with the sound file or the image file 110 with the identified audio 132 to generate a processing result 122, and then automatically performs the processing of the sound file or image file 110 according to the processing result 122. Classification.

Various examples of related audio related to the sound file or the video file 110 will be described below with reference to FIG. 3 and FIG. The document classification system 300 in FIG. 3 is substantially the same as the document classification system 200 in FIG. 2 except that an audio capture device 360 is added. In practice, the audio capture device 360 can be a sound card, an audio processing chip or other similar components.

After the audio file or image file 110 is recorded, the audio capture device 360 retrieves a pending audio signal 362 extracted from the audio file or image file 110 as the relevant audio associated with the sound file or the image file 110 . Next, the processor 120 compares the pending audio 362 with the recognized audio 132 to generate a processing result 122 , and then automatically classifies the audio file or the video file 110 according to the processing result 122 .

Regarding the specific classification method, in one embodiment, before audio recording or video recording, the lecturer can operate the file classification system 300 or an external computer device to customize the exclusive category 334 (such as: a file folder of the Windows operating system), so that The storage device 130 records the category 334, and the processor 120 establishes the association between the path of the category 334 and the identified audio 132. After the sound file or image file 110 is recorded, the processor 120 analyzes and compares the acoustics of the audio 362 to be processed. The acoustic feature is the acoustic feature of the identified audio 132 . When the acoustic feature of the pending audio 362 matches the acoustic feature of the identified audio 132 , the processor 120 classifies the audio file or image file 110 into the category 234 .

Or, in another embodiment, after the sound file or image file 110 is recorded, the processor 120 analyzes and compares the semantic feature (semantic feature) of the pending audio 362 with the semantic feature of the identification audio 132, when the pending audio When the semantic feature of 362 matches the semantic feature of the recognized audio 132 , the processor 120 classifies the audio file or video file 110 into the category 234 .

In the speech recognition mechanism of the document classification system 300, the above-mentioned acoustic features such as the composition ratio of the human voice and the background sound can be used to assist in judging the scene recognition or human voice recognition; the scene recognition can use the nature of the background sound to infer the surrounding objects, events; human voice recognition can use the sound quality characteristics of the human voice, such as voiceprint, as a basis for comparison. Semantic features can be used, for example, to identify key words in the audio, or through commonly used words, names, and so on. The above-mentioned acoustic features and semantic features are not limited to the enumerated scope, and any acoustic feature or semantic feature that can be used as a basis for scene recognition shall be included within the scope of the present invention.

In addition to the pending audio 362 can be used as related audio related to the sound file or the image file 110, another example of related audio, please refer to FIG. 4, which is a file classification system 400 according to the fourth embodiment of the present invention . The hardware architecture of the document classification system 400 in FIG. 4 is substantially the same as the hardware architecture of the document classification system 200 in FIG. 2 .

After the sound file or image file 110 is recorded, the speaker reads a sentence as supplementary audio 452, so that the audio recording device 250 receives the supplementary audio 452 as the above-mentioned relevant audio related to the sound file or the image file 110, and then processes The device 120 compares the supplementary audio 452 with the recognized audio 132 to generate a processing result 122 , and then automatically classifies the audio file or the video file 110 according to the processing result 122 .

Regarding the specific classification method, in one embodiment, before audio recording or video recording, the lecturer can operate the file classification system 400 or an external computer device to customize the exclusive category 434 (such as: a file folder of the Windows operating system), so that The storage device 130 records the category 434, and the processor 120 establishes the association between the path of the category 434 and the identified audio 132. After the sound file or image file 110 is recorded, the processor 120 analyzes and compares the acoustics of the supplementary audio 452. The feature matches the acoustic feature of the identified audio 132 , and the processor 120 classifies the sound file or video file 110 into the category 434 when the acoustic feature of the supplemental audio 452 matches the acoustic feature of the identified audio 132 .

Or, in another embodiment, after the audio file or image file 110 is recorded, the processor 120 analyzes and compares the semantic features of the supplementary audio 452 with the semantic features of the identification audio 132, when the semantic features of the supplementary audio 452 are in line with The processor 120 classifies the audio file or the video file 110 into a category 434 when the semantic features of the identified audio 132 match.

In the voice recognition mechanism of the document classification system 400, acoustic features such as the composition ratio of the human voice and the background sound can be used to assist in the judgment of scene recognition or human voice recognition; the scene recognition can predict the surrounding objects and occurrences through the nature of the background sound. events; human voice recognition can use the sound quality characteristics of the human voice, such as voiceprint, as a basis for comparison. Semantic features can be used, for example, to identify key words in the audio, or through commonly used words, names, and so on. The above-mentioned acoustic features and semantic features are not limited to the enumerated scope, and any acoustic feature or semantic feature that can be used as a basis for scene recognition shall be included within the scope of the present invention.

FIG. 5 is a flowchart of a file classification method 500 according to the fifth embodiment of the present invention. The document classification method 500 can be implemented through a computer system, such as the aforementioned document classification systems 100, 200, 300, 400, and part of the functions can also be implemented as at least one computer program and stored in a computer-readable recording medium Among them, the at least one computer program has a plurality of instructions, and these instructions cause the computer to execute the file classification method 500 when executed on a computer.

As shown in FIG. 5 , the document classification method 500 includes a plurality of steps S502-S506. However, those skilled in the present case should understand that the steps mentioned in this embodiment, unless the order is specifically stated, can be adjusted according to actual needs, and can even be executed simultaneously or partly simultaneously. As for the hardware device for implementing these steps, since the above embodiments have been specifically disclosed, it is not repeated here.

Firstly, before recording or video recording, in step S502, the recognition audio is pre-recorded and stored as a reference sample for speech recognition. The recognition audio can be a segment of speech of a user (eg, a speaker). Then, after the recording or video recording is completed, in step S504, the audio file or image file is received, the system can prompt the request or the lecturer selects the voice category, and then, in step S506, a related audio file or the image file will be linked The audio is compared with the recognized audio to generate a processing result, and then automatically classify the sound file or image file according to the processing result. In this way, through the voice file recognition mechanism of the file classification method 500, the files can be classified quickly without cumbersome manual operations.

Specifically, before audio recording or video recording, the lecturer can operate the computer system to customize an exclusive category (such as: a file folder of the Windows operating system), and in step S502, establish the association between the path of the category and the identified audio . In step S506, if the relevant audio associated with the sound file or the image file matches the identified audio, it means that it is the audio or video recording of the speaker himself. Therefore, the processor 120 automatically classifies the sound file or the image file 110 to the category defined by the lecturer in the storage device 130 .

There are at least two ways to implement the above-mentioned related audio. Regarding the first way, in one embodiment, step S506 extracts a pending audio from a sound file or an image file as a related audio, analyzes and compares the pending audio The acoustic feature matches at least one acoustic feature for identifying audio, and when the acoustic feature of the pending audio matches the at least one acoustic feature for identifying audio, the audio file or image file is classified into the category.

Or, in another embodiment, step S506 analyzes and compares the semantic features of the pending audio with the semantic features of at least one recognized audio, and when the semantic features of the pending audio match the semantic features of at least one recognized audio, the sound Documents or image files are classified into this category.

Regarding the second implementation of the related audio, in one embodiment, step S506 receives a supplementary audio as the related audio, where the supplementary audio can be a sentence read by the speaker after the recording or video recording is completed; then, Step S506 analyzes and compares at least one acoustic feature of the identified audio and the acoustic feature of the supplementary audio, and when at least one acoustic feature of the identified audio matches the acoustic feature of the supplementary audio, the audio file or video file is classified into the category.

Or, in another embodiment, step S506 analyzes and compares at least one semantic feature of the recognized audio with the semantic feature of the supplementary audio, and when at least one semantic feature of the recognized audio matches the semantic feature of the supplementary audio, the pending Audio falls into this category.

When the above-mentioned step S506 analyzes and compares the acoustic features, the audio frequency, frequency spectrum, amplitude, phase, duration, voice print or any other audio frequency can also be used. Combination, or the result of mathematical operation, time domain to frequency domain transform (time domain to frequency domain transform) results, as the basis for analysis and comparison, and should be included in the patent scope of the present invention. When performing scene recognition, you can recognize surrounding objects, such as the footsteps of different types of shoes, different vehicles, and the sounds of different animals; you can also recognize events that occur, such as wind, rain, door opening and closing, Different types of music, etc.; human voice recognition can identify the uniqueness of different people according to their pitch, accent, rhythm, volume, tone quality, etc. sex. Alternatively, the user can define the strength of the matching degree of the acoustic features and the order of the judgment conditions according to the requirements, and store them in the storage device. All acoustic features that can be used as the basis for scene recognition or human voice recognition should be included in the patent scope of the present invention.

In the above step S506, when analyzing the semantic features, the method includes identifying key words in the audio, such as the terms of different sports, and the terms of different occasions (speech, wedding ceremony, graduation ceremony, concert, etc.) can be used for scene recognition; if For human voice recognition, common words, names, and related terms can be used. The above matching degree is exemplified as follows: hypernym and hyponym, synonyms, similar concepts, translation vocabulary in different languages, names or part of names in different languages can all be defined as high matching degree, and can be The strength of the matching degree and the sequence of the judgment conditions are defined according to the needs of the user, and stored in the storage device.

The above-mentioned step S506 can be set to automatically classify according to the processing result, or to allow users to confirm the processing result by suggesting classification into a category, which can be flexibly adjusted according to the needs of use; an example of one of the classification forms is to file to For example, in a folder of the Windows operating system.

To sum up, the present invention can classify the files into appropriate categories in real time after each recording or video recording through the above-mentioned embodiment, so as to solve the problem that the recording or video files have not been sorted out for a long time or there are a large number of files , it is difficult to quickly find a specific file, and it also avoids the trouble of manual file classification.

Although the present invention has been disclosed above in terms of implementation, it is not intended to limit the present invention. Any skilled person can make various changes and modifications without departing from the spirit and scope of the present invention. Therefore, the present invention The scope of protection should depend on the scope defined by the claims.

Claims (16)

1. a document classification system, is characterized in that, comprises:

One storage device, stores at least one identification audio frequency;

One receiving trap, obtains an audio files or an image file; And

One processor, the related audio and this at least one identification audio frequency that pass are connected in this audio files or this image file compare to produce a result, and then automatically carry out the classification of this audio files or this image file according to this result.

2. document classification system according to claim 1, is characterized in that, separately comprise:

One voice recorder, before this audio files or this image file are recorded, this at least one identification audio frequency of pre-recording.

3. document classification system according to claim 2, is characterized in that, separately comprise:

One audio frequency capture device, captures an audio frequency co-pending using as this related audio from this audio files or this image file.

4. document classification system according to claim 3, is characterized in that, it is characterized in that, at least one classification of this storage device record, is set up the path of this at least one classification and associating between this at least one identification audio frequency by this processor,

After this audio files or this image file are completed by recording, this processor is analyzed and the acoustic feature of this audio frequency co-pending of comparison and the acoustic feature of this at least one identification audio frequency, when the acoustic feature of this audio frequency co-pending and the acoustic feature of this at least one identification audio frequency match, this audio files or this image file are sorted out to this at least one classification by this processor.

5. document classification system according to claim 3, is characterized in that, at least one classification of this storage device record, is set up the path of this at least one classification and associating between this at least one identification audio frequency by this processor,

After this audio files or this image file are completed by recording, this processor is analyzed and the meaning of one's words feature of this audio frequency co-pending of comparison and the meaning of one's words feature of this at least one identification audio frequency, when the meaning of one's words feature of this audio frequency co-pending and the meaning of one's words feature of this at least one identification audio frequency match, this audio files or this image file are sorted out to this at least one classification by this processor.

6. document classification system according to claim 1, is characterized in that, separately comprise:

One voice recorder, after this audio files or this image file are completed by recording, receives a supplementary audio using as this related audio.

7. document classification system according to claim 6, is characterized in that, at least one classification of this storage device record, is set up the path of this at least one classification and associating between this at least one identification audio frequency by this processor,

After this audio files or this image file are completed by recording, this processor is analyzed and comparison this at least one acoustic feature of identification audio frequency and acoustic feature of this supplementary audio, when this at least one acoustic feature of identification audio frequency and the acoustic feature of this supplementary audio match, this audio files or this image file are sorted out to this at least one classification by this processor.

8. document classification system according to claim 6, is characterized in that, at least one classification of this storage device record, is set up the path of this at least one classification and associating between this at least one identification audio frequency by this processor,

After this audio files or this image file are completed by recording, this processor is analyzed and the meaning of one's words feature of this at least one identification audio frequency of comparison and the meaning of one's words feature of this supplementary audio, when the meaning of one's words feature of this at least one identification audio frequency and the meaning of one's words feature of this supplementary audio match, this audio files or this image file are sorted out to this at least one classification by this processor.

9. a file classifying method, is characterized in that, comprises the following step:

(A) at least one identification audio frequency is stored;

(B) audio files or an image file is obtained; And

(C) related audio and this at least one identification audio frequency that pass are connected in this audio files or this image file compare to produce a result, and then automatically carry out the classification of this audio files or this image file according to this result.

10. file classifying method according to claim 9, is characterized in that, step (A) comprises:

Before this audio files or this image file are recorded, this at least one identification audio frequency of pre-recording.

11. file classifying methods according to claim 10, is characterized in that, step (C) comprises:

An audio frequency co-pending is captured using as this related audio from this audio files or this image file.

12. file classifying methods according to claim 11, is characterized in that, step (A) comprises: set up the path of at least one classification and associating between this at least one identification audio frequency; Step (C) separately comprises: after this audio files or this image file are completed by recording, analyze and the acoustic feature of this audio frequency co-pending of comparison and the acoustic feature of this at least one identification audio frequency, when the acoustic feature of this audio frequency co-pending and the acoustic feature of this at least one identification audio frequency match, this audio files or this image file are sorted out to this at least one classification.

13. file classifying methods according to claim 11, is characterized in that, step (A) comprises: set up the path of at least one classification and associating between this at least one identification audio frequency; Step (C) separately comprises: after this audio files or this image file are completed by recording, analyze and the meaning of one's words feature of this audio frequency co-pending of comparison and the meaning of one's words feature of this at least one identification audio frequency, when the meaning of one's words feature of this audio frequency co-pending and the meaning of one's words feature of this at least one identification audio frequency match, this audio files or this image file are sorted out to this at least one classification.

14. file classifying methods according to claim 9, is characterized in that, step (C) comprises:

After this audio files or this image file are completed by recording, receive a supplementary audio using as this related audio.

15. file classifying methods according to claim 14, is characterized in that, step (A) comprises: set up the path of at least one classification and associating between this at least one identification audio frequency; Step (C) separately comprises: after this audio files or this image file are completed by recording, analyze and comparison this at least one acoustic feature of identification audio frequency and acoustic feature of this supplementary audio, when this at least one acoustic feature of identification audio frequency and the acoustic feature of this supplementary audio match, this audio files or this image file are sorted out to this at least one classification.

16. file classifying methods according to claim 14, is characterized in that, step (A) comprises: set up the path of at least one classification and associating between this at least one identification audio frequency; Step (C) separately comprises: after this audio files or this image file are completed by recording, analyze and the meaning of one's words feature of this at least one identification audio frequency of comparison and the meaning of one's words feature of this supplementary audio, when the meaning of one's words feature of this at least one identification audio frequency and the meaning of one's words feature of this supplementary audio match, this audio frequency co-pending is sorted out to this at least one classification.