CN104751033B - A kind of user behavior authentication method and device based on audio-video document - Google Patents

Abstract

This application provides a user behavior authentication method and device based on audio and video files. The method includes: judging whether the video of the audio-video file uploaded by the user satisfies the authentication condition; if the video of the audio-video file satisfies the authentication condition, converting the audio of the audio-video file into the corresponding text to be authenticated; The user behavior is authenticated according to the text to be authenticated corresponding to the audio and video files. This application realizes automatic authentication of user behavior based on audio and video files, saves human resources, and improves authentication efficiency.

Description

A user behavior authentication method and device based on audio and video files

technical field

The present application relates to the field of computer technology, in particular to a method and device for user behavior authentication based on audio and video files.

Background technique

With the extensive development of Internet technology, users can implement various business operations through the Internet. For example: users can communicate through the Internet, and users can also conduct financial loans through the Internet. To realize the above-mentioned business, the service provider often needs to authenticate the user's behavior first, for example: through the user's behavior to verify whether the user has clearly understood the regulations of business operations, etc.

In related technologies, a user can upload an audio and video file, and then the service provider authenticates the user's behavior based on the audio and video file. However, at present, service providers usually use manual authentication, which has low efficiency and high cost. Therefore, an automatic user behavior authentication scheme based on audio and video files is urgently needed.

Contents of the invention

In view of this, the present application provides a user behavior authentication method and device based on audio and video files.

Specifically, this application is achieved through the following technical solutions:

A user behavior authentication method based on audio and video files, said method comprising:

Determine whether the video of the audio and video files uploaded by the user meets the authentication conditions;

If the video of the audio-video file satisfies the authentication condition, then convert the audio of the audio-video file into corresponding text to be authenticated;

The user behavior is authenticated according to the text to be authenticated corresponding to the audio and video files.

Further, the judging whether the video of the audio-video file uploaded by the user satisfies the authentication condition includes:

Extract multiple pictures from the audio and video files;

Judging whether the similarity of the backgrounds of the plurality of pictures is greater than or equal to a first threshold;

If the similarity between the backgrounds of the multiple pictures is greater than or equal to the first threshold, it is confirmed that the video of the audio and video file satisfies the authentication condition.

Further, converting the audio of the audio-video file into corresponding text to be authenticated includes:

Divide the audio of the audio-video file into N audio segments according to the preset first segmentation rule, where N is a natural number greater than 1;

converting the audio in the audio segment into corresponding initial text;

The N initial texts corresponding to the N audio segments are combined to obtain the text to be authenticated corresponding to the audio and video files.

Further, after identifying the audio in the audio clip as the corresponding initial text, the method further includes:

Extracting N-1 verification audio segments from the audio of the audio and video file according to the preset second segmentation rule;

converting the audio in the verification audio segment into corresponding verification text;

Judging whether the N initial texts corresponding to the N audio segments are accurate according to the N-1 verification texts;

If the N initial texts corresponding to the N audio segments are accurate, combine the N initial texts corresponding to the N audio segments.

Further, the authentication of user behavior according to the text to be authenticated corresponding to the audio and video files includes:

Calculating the text similarity between the text to be authenticated corresponding to the audio and video file and the preset authentication text;

When the text similarity is greater than or equal to the preset second threshold, it is confirmed that the user has passed the authentication.

A user behavior authentication device based on audio and video files, said device comprising:

A judging unit, configured to judge whether the video of the audio-video file uploaded by the user satisfies the authentication condition;

A converting unit, configured to convert the audio of the audio and video file into corresponding text to be authenticated when the video of the audio and video file satisfies the authentication condition;

The authentication unit is configured to authenticate the user behavior according to the text to be authenticated corresponding to the audio and video file.

Further, the judging unit is specifically configured to extract a plurality of pictures from the audio and video files, judge whether the background similarity of the plurality of pictures is greater than or equal to a first threshold, and check whether the background similarity of the plurality of pictures is greater than or equal to a first threshold When the similarity degree is greater than or equal to the first threshold, it is confirmed that the video of the audio and video file satisfies the authentication condition.

Further, the conversion unit is specifically configured to divide the audio of the audio and video file into N audio segments according to a preset first segmentation rule, where N is a natural number greater than 1; convert the audio in the audio segment is the corresponding initial text; combine the N initial texts corresponding to the N audio segments to obtain the text to be authenticated corresponding to the audio and video file.

Further, the device also includes:

A verification unit, configured to extract N-1 verification audio segments from the audio of the audio and video file according to a preset second segmentation rule after identifying the audio in the audio segment as the corresponding initial text ; Convert the audio in the verification audio segment into corresponding verification text; judge whether the N initial texts corresponding to the N audio segments are accurate according to the N-1 verification texts;

The conversion unit specifically combines the N initial texts corresponding to the N audio segments when the N initial texts corresponding to the N audio segments are accurate.

Further, the authentication unit specifically calculates the text similarity between the text to be authenticated corresponding to the audio and video file and the preset authentication text, and when the text similarity is greater than or equal to the preset second threshold, confirms that the user Certification passed.

It can be seen from the above description that after confirming that the video of the audio and video file uploaded by the user satisfies the authentication conditions, the application can convert the audio of the audio and video file into the corresponding text to be authenticated, and provide the user with the text according to the text to be authenticated. Behavior authentication, so that user behavior can be automatically authenticated based on audio and video files, saving human resources and improving authentication efficiency.

Description of drawings

Fig. 1 is a flowchart of a user behavior authentication method based on audio and video files in an exemplary embodiment of the present application.

Fig. 2 is a flowchart of another user behavior authentication method based on audio and video files in an exemplary embodiment of the present application.

Fig. 3 is a schematic structural diagram of a terminal in an exemplary embodiment of the present application.

Fig. 4 is a schematic structural diagram of a user behavior authentication device based on audio and video files in an exemplary embodiment of the present application.

detailed description

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with aspects of the present application as recited in the appended claims.

The terminology used in this application is for the purpose of describing particular embodiments only, and is not intended to limit the application. As used in this application and the appended claims, the singular forms "a", "the", and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It should also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

It should be understood that although the terms first, second, third, etc. may be used in this application to describe various information, the information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of the present application, first information may also be called second information, and similarly, second information may also be called first information. Depending on the context, the word "if" as used herein may be interpreted as "at" or "when" or "in response to a determination."

In view of the above problems, the present application provides an automatic authentication scheme based on user behavior of audio and video files.

Fig. 1 is a flowchart of a user behavior authentication method based on audio and video files in an exemplary embodiment of the present application.

Please refer to Fig. 1, the application provides a user behavior authentication method based on audio and video files, the method can be applied on the terminal, including the following steps:

Step 101, judging whether the video of the audio and video file uploaded by the user satisfies the authentication condition.

In this embodiment, multiple pictures can be randomly extracted from the audio and video files uploaded by the user, and then it is judged whether the background similarity of the multiple pictures is greater than or equal to the preset first threshold. When the similarity of the background is greater than or equal to the first threshold, it can be confirmed that the video of the audio and video file satisfies the authentication condition.

Step 102, if the video of the audio-video file satisfies the authentication condition, convert the audio of the audio-video file into corresponding text to be authenticated.

In this embodiment, when the video of the video and audio file satisfies the authentication condition, the audio of the audio and video file may be converted into the corresponding text to be authenticated. Specifically, the audio may be converted into corresponding text to be authenticated by a voice recognition method. In order to improve the accuracy of audio conversion, the audio can also be divided into a plurality of audio segments, and then the audio of each audio segment is converted into a corresponding initial text, and the audio corresponding to the audio is obtained by combining each initial text. Authentication text. Of course, those skilled in the art can also use other methods to convert the audio of the audio-video file into the corresponding text to be authenticated, which is not particularly limited in this application.

Step 103: Authenticate the user behavior according to the text to be authenticated corresponding to the audio/video file.

In this embodiment, the text to be authenticated corresponding to the audio and video file can be matched with the preset authentication text, such as: calculating the text similarity between the text to be authenticated and the authentication text, when the text similarity is greater than When it is equal to the second threshold, it is confirmed that the user authentication is passed.

It can be seen from the above description that after confirming that the video of the audio and video file uploaded by the user satisfies the authentication conditions, the application can convert the audio of the audio and video file into the corresponding text to be authenticated, and provide the user with the text according to the text to be authenticated. Behavior authentication, so that user behavior can be automatically authenticated based on audio and video files, saving human resources and improving authentication efficiency.

Fig. 2 is a flowchart of another user behavior authentication method based on audio and video files in an exemplary embodiment of the present application.

Please refer to Figure 2, this application provides another user behavior authentication method based on audio and video files, the method can be applied to the terminal, including the following steps:

Step 201, extract multiple pictures from the audio and video files uploaded by the user.

In this embodiment, when the service provider wants to verify whether the user has clearly understood the regulations of the business operation, the user can upload an audio and video file that reads the contents of the business operation regulations aloud, so as to provide the basis for authentication. Generally speaking, if the audio and video file uploaded by the user is an audio and video file specified by the user to read aloud the business operation, the shooting background of the audio and video file remains unchanged or changes little. Therefore, the background of the audio and video files can be judged first.

Specifically, in this step, multiple pictures may be first extracted from the audio and video files. For example: multiple pictures may be extracted from the video frames of the audio and video files through a random algorithm. The number of pictures to be extracted can be set by the developer, and this application does not impose any special limitation on this.

Step 202, judging whether the similarity of the backgrounds of the multiple pictures is greater than or equal to a first threshold, and if the similarity of the backgrounds of the multiple pictures is greater than or equal to the first threshold, perform step 203.

Based on the aforementioned step 201, after multiple pictures are extracted, face recognition may be performed on each picture, and then the face in the picture is removed to obtain the background of the picture. After obtaining the multiple backgrounds of the multiple pictures, calculate the similarity of the backgrounds of the multiple pictures, such as: SIFT (Scale-invariant feature transform, scale-invariant feature transformation) algorithm, MD5 (message-digest algorithm5 , Information Digest) algorithm and other algorithms to calculate the similarity of the backgrounds of the multiple pictures, which is not particularly limited in the present application.

Specifically, in this embodiment, for the plurality of pictures, the similarity of the background of any two pictures can be calculated, such as: the similarity S _ab of picture a and picture b, and then the calculated multiple similarities S The average value of _ab , using the average value as the similarity of the background of the plurality of pictures.

In this step, it is judged whether the similarity of the backgrounds of the plurality of pictures is greater than or equal to the first threshold. Wherein, the first threshold may be set by a developer, and this application does not make any special limitation on this. If the similarity of the backgrounds of the multiple pictures is greater than or equal to the first threshold, it can be confirmed that the shooting background of the audio and video files does not change or changes very little, and the audio and video files meet the authentication conditions, and the steps are executed 203. If the similarity of the backgrounds of the multiple pictures is smaller than the first threshold, step 201 may be re-executed to re-extract multiple pictures for determination. When the preset number of re-extractions is reached, if the similarity of the backgrounds of the extracted pictures is still less than the first threshold, it can be confirmed that the video of the audio and video files does not meet the authentication conditions, such as: the user casually finds An audio and video file is uploaded. In this embodiment, an authentication failure message may be returned.

Step 203, confirming that the video of the audio/video file satisfies the authentication condition.

Based on the judgment result of step 202, when the background similarity of the multiple pictures is greater than or equal to the first threshold, it can be confirmed that the video of the audio/video file satisfies the authentication condition, and step 204 is executed.

Optionally, in another embodiment of the present application, after extracting multiple pictures from the audio and video files uploaded by the user, it is also possible to perform face recognition on each picture, and then determine the people of the multiple pictures. Face similarity, when the face similarity of the multiple pictures is greater than or equal to a certain preset threshold, it is confirmed that the video of the audio and video file satisfies the authentication condition, and step 204 is executed.

Step 204: Divide the audio of the audio and video file into N audio segments according to a preset first division rule.

In this embodiment, when it is confirmed that the video of the audio-video file satisfies the authentication condition, the audio of the audio-video file is converted into the corresponding text to be authenticated, so as to complete the authentication process. In actual implementation, for the audio of the audio and video file, denoising processing can be performed first, such as: the interference in the audio can be eliminated by analyzing the frequency spectrum, and then the audio and video can be divided according to the preset first segmentation rule The audio of the file is divided into N audio segments.

Specifically, in this embodiment, the audio of the audio and video file can be divided into N audio segments according to a preset first segmentation rule, wherein the value of N is a natural number greater than 1, and the first A division rule can be set by the developer, for example: the audio of the audio/video file can be divided into audio segments with a duration of 5 seconds according to the order of time.

Step 205, converting the audio in the audio segment into corresponding initial text.

Based on the aforementioned step 204, after obtaining the N audio judgments of the audio in the audio and video clips, for each audio clip, the audio clip can be converted into a corresponding text according to the speech recognition method provided in the related art. In this embodiment, the text corresponding to the audio segment is called the initial text.

Step 206, combine the N initial texts corresponding to the N audio clips to obtain the text to be authenticated corresponding to the audio and video files.

Based on the aforementioned step 205, after obtaining the N initial texts corresponding to the N audio clips, in this step, the N initial files are combined in chronological order to obtain the corresponding audio and video files to be authenticated. text.

Further, in this embodiment, in order to improve the accuracy of the text to be authenticated, after the audio in the audio clip is identified as the corresponding initial text in step 205, a verification of the accuracy of the initial text is performed process. For example: N-1 verification audio segments can be extracted from the audio of the audio-video file according to the preset second segmentation rule, and then the audio in the verification audio segments is converted into a corresponding verification text, And judge whether the N initial texts corresponding to the N audio segments are accurate according to the N-1 verification texts, and if accurate, combine the N initial texts corresponding to the N audio segments to obtain the audio The text to be authenticated corresponding to the video file. The second segmentation rule can also be set by the developer. For example, each of the N-1 verification audio segments extracted according to the second segmentation rule is the same as that of each verification audio segment according to the first Two consecutive audio segments divided by a segmentation rule have an intersection. Suppose, according to the first segmentation rule, the audio of the audio and video file uploaded by the user is divided into three audio segments, wherein the time information of the first audio segment is 0-5 seconds of the audio, and the time information of the second audio segment The information is 5-10 seconds of the audio, and the time information of the third audio segment is 10-15 seconds of the audio. Then the 2 verification audio segments that can be extracted according to the second segmentation rule, wherein, the time information of the first verification audio segment can be 4-7 seconds of the audio, and the first verification audio segment and The first audio clip and the second audio clip have an intersection, and the time information of the second verification audio clip is 9-12 seconds of the audio, and the second verification audio clip and the second audio clip have an intersection. segment and said third audio segment have an intersection.

In this embodiment, after extracting the N-1 verification audio segments, the audio in the verification audio segments is converted into corresponding verification text, and then the verification text corresponding to the verification audio segments is determined. Whether the verification text is included in the initial text corresponding to the audio segment intersecting with the verification audio segment, if yes, it can be confirmed that the N initial texts corresponding to the N audio segments are accurate. Still taking the aforementioned assumptions as an example, if the initial text corresponding to the aforementioned first audio segment is "the applicant has read", the initial text corresponding to the aforementioned second audio segment is "the above rules stipulate", and the initial text corresponding to the aforementioned third audio segment is "February 1, 2015", and the verification text corresponding to the aforementioned first verification audio segment is "read the above", and the verification text corresponding to the aforementioned second verification audio segment is "stipulation, 2015", then The verification text "read the above" corresponding to the first verification audio segment contains the initial text "the applicant has read" corresponding to the first audio segment and the initial text "the above rule stipulates" corresponding to the second audio segment In ", the verification text corresponding to the second verification audio segment "stipulations, 2015" contains the initial text corresponding to the second audio segment "the above rules stipulate" and the initial text corresponding to the third audio segment In "February 1, 2015", it can be confirmed that the initial text corresponding to the aforementioned three audio clips is accurate. If the verification text corresponding to the verification audio segment is not included in the initial text corresponding to the audio segment that the verification audio segment intersects with, it can be confirmed that the N initial texts corresponding to the N audio segments are inaccurate.

In this embodiment, when it is confirmed that the N initial texts corresponding to the N audio segments are inaccurate, N-1 verification audio segments can be re-extracted for verification according to the preset third segmentation rule. The result is still inaccurate, then the initial text can be amended according to the verification text corresponding to the N-1 verification audio segments, or the audio and video files uploaded by the user can be resubmitted according to the preset fourth segmentation rule. The audio is divided into M audio clips, and then step 205 and the subsequent verification process are executed, which is not particularly limited in this application.

Step 207, calculating the text similarity between the text to be authenticated corresponding to the audio and video file and the preset authentication text.

Based on the aforementioned step 206, after the text to be authenticated corresponding to the audio and video file is obtained, the text similarity between the text to be authenticated and the preset authentication text is calculated. Specifically, in this step, the text similarity between the text to be authenticated and the preset authentication text can be calculated according to the text similarity algorithm provided in the related art, and the present application will not repeat them here.

Step 208, when the text similarity is greater than or equal to the preset second threshold, confirm that the user has passed the authentication.

Based on the aforementioned step 207, after calculating the text similarity between the text to be authenticated and the preset authentication text, it is judged whether the text similarity is greater than or equal to a second threshold, and the second threshold can be set by the developer. If the text similarity is greater than or equal to a preset second threshold, it can be confirmed that the user authentication has passed, and if the text similarity is smaller than the second threshold, it can be confirmed that the user authentication has failed.

Further, in another optional embodiment of the present application, a third threshold may also be set, and the third threshold is smaller than the second threshold. When the text similarity is less than the second threshold, it can also be judged whether the text similarity is greater than or equal to the third threshold, and if the text similarity is greater than or equal to the third threshold, a prompt can be output to the administrator to Remind management personnel to perform manual verification of user behavior based on the audio and video files.

It can be seen from the above description that after confirming that the video of the audio and video file uploaded by the user satisfies the authentication conditions, the application can convert the audio of the audio and video file into the corresponding text to be authenticated, and provide the user with the text according to the text to be authenticated. Behavior authentication, so that user behavior can be automatically authenticated based on audio and video files, saving human resources and improving authentication efficiency.

Corresponding to the aforementioned embodiments of the audio-video file-based user behavior authentication method, the present disclosure also provides an embodiment of an audio-video file-based user behavior authentication device.

Corresponding to the embodiments of the user behavior authentication method based on audio and video files in this application, this application also provides a user behavior authentication device based on audio and video files. The devices described in this application can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, the user behavior authentication device based on audio and video files in this application is a logical device, which reads the corresponding computer program instructions in the non-volatile memory into the memory through the processor of the device where it is located formed by running.

Please refer to FIG. 3 and FIG. 4 , the present application provides a user behavior authentication device 300 based on audio and video files.

Wherein, the judging unit 301 is used to judge whether the video of the audio-video file uploaded by the user satisfies the authentication condition;

The converting unit 302 is configured to convert the audio of the audio and video file into corresponding text to be authenticated when the video of the audio and video file satisfies the authentication condition;

The authentication unit 303 is configured to authenticate user actions according to the text to be authenticated corresponding to the audio and video files.

Further, the judging unit 301 is specifically configured to extract a plurality of pictures from the audio and video files, judge whether the background similarity of the plurality of pictures is greater than or equal to a first threshold, and determine whether When the similarity of the background is greater than or equal to the first threshold, it is confirmed that the video of the audio and video file satisfies the authentication condition.

Further, the conversion unit 302 is specifically configured to divide the audio of the audio and video file into N audio segments according to a preset first segmentation rule, where N is a natural number greater than 1; the audio in the audio segment Converting to corresponding initial texts; combining the N initial texts corresponding to the N audio clips to obtain the text to be authenticated corresponding to the audio and video files.

The checking unit 304 is configured to extract N-1 checksums from the audio of the audio and video file according to the preset second segmentation rule after identifying the audio in the audio clip as the corresponding initial text. Check the audio clip; convert the audio in the verification audio clip into a corresponding verification text; judge whether the N initial texts corresponding to the N audio clips are accurate according to the N-1 verification texts;

The conversion unit specifically combines the N initial texts corresponding to the N audio segments when the N initial texts corresponding to the N audio segments are accurate.

Further, the authentication unit 303 specifically calculates the text similarity between the text to be authenticated corresponding to the audio and video file and the preset authentication text, and when the text similarity is greater than or equal to the preset second threshold, confirms User authentication passed.

For the implementation process of the functions and effects of each unit in the above device, please refer to the implementation process of the corresponding steps in the above method for details, and will not be repeated here.

The above is only a preferred embodiment of the application, and is not intended to limit the application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the application should be included in the application. within the scope of protection.

Claims (8)

1. a kind of user behavior authentication method based on audio-video document, it is characterised in that methods described includes：

Judge whether the video for the audio-video document that user uploads meets authentication condition；

If the video of the audio-video document meets the authentication condition, the audio of the audio-video document is converted to pair The text to be certified answered；

Text to be certified is authenticated to user behavior according to corresponding to the audio-video document；

Whether the video of the audio-video document for judging user and uploading meets authentication condition, including：

Plurality of pictures is extracted from the audio-video document；

Judge whether the similarity of the background of the plurality of pictures is more than or equal to first threshold；

If the similarity of the background of the plurality of pictures is more than or equal to the first threshold, the audio-video document is confirmed Video meets the authentication condition.

2. according to the method for claim 1, it is characterised in that the audio by the audio-video document is converted to correspondingly Text to be certified, including：

The audio of the audio-video document is divided into by N number of audio fragment according to the default first segmentation rule, N is oneself more than 1 So number；

Audio in the audio fragment is converted into corresponding original text；

N number of original text corresponding to N number of audio fragment is combined, to obtain text to be certified corresponding to the audio-video document This.

3. according to the method for claim 2, it is characterised in that be corresponding by the audio identification in the audio fragment After original text, methods described also includes：

N-1 verification audio fragment is extracted from the audio of the audio-video document according to the default second segmentation rule；

Audio in the verification audio fragment is converted into corresponding verification text；

Whether N number of original text according to corresponding to described N-1 verification text judges N number of audio fragment is accurate；

If N number of original text corresponding to N number of audio fragment is accurate, combine N number of first corresponding to N number of audio fragment Beginning text.

4. according to the method for claim 1, it is characterised in that the text to be certified according to corresponding to the audio-video document This is authenticated to user behavior, including：

Calculate the text similarity of text to be certified corresponding to the audio-video document and default certification text；

When the text similarity is more than or equal to default Second Threshold, confirm that user authentication passes through.

5. a kind of user behavior authentication device based on audio-video document, it is characterised in that described device includes：

Whether judging unit, the video of the audio-video document uploaded for judging user meet authentication condition；

Converting unit, for when the video of the audio-video document meets the authentication condition, by the audio-video document Audio is converted to corresponding text to be certified；

Authentication unit, user behavior is authenticated for the text to be certified according to corresponding to the audio-video document；

The judging unit, specifically for extracting plurality of pictures from the audio-video document, judge the back of the body of the plurality of pictures Whether the similarity of scape is more than or equal to first threshold, and the similarity of the background in the plurality of pictures is more than or equal to described first During threshold value, confirm that the video of the audio-video document meets the authentication condition.

6. device according to claim 5, it is characterised in that

The converting unit, it is N number of specifically for being divided into the audio of the audio-video document according to the default first segmentation rule Audio fragment, N are the natural number more than 1；Audio in the audio fragment is converted into corresponding original text；Described in combination N number of original text corresponding to N number of audio fragment, to obtain text to be certified corresponding to the audio-video document.

7. device according to claim 6, it is characterised in that described device also includes：

Verification unit, for after by the audio identification in the audio fragment being corresponding original text, according to default Second segmentation rule extracts N-1 verification audio fragment from the audio of the audio-video document；By the verification audio piece Audio in section verifies text corresponding to being converted to；According to corresponding to described N-1 verification text judges N number of audio fragment Whether N number of original text is accurate；

The converting unit, specifically when N number of original text is accurate corresponding to N number of audio fragment, combine N number of audio N number of original text corresponding to fragment.

8. device according to claim 5, it is characterised in that

The authentication unit, specifically calculate the text of text to be certified corresponding to the audio-video document and default certification text Similarity, and when the text similarity is more than or equal to default Second Threshold, confirm that user authentication passes through.