Voiceprint recognition testing method, device, storage medium and electronic device

Abstract

The present application discloses a voiceprint recognition test method, device, storage medium and electronic device, which relates to the field of artificial intelligence technology. The method comprises: obtaining an image frame in a target video stream, wherein the image frame includes text information; extracting the text information in the image frame; using the text information to match in a preset voice library to obtain the voice voiceprint information corresponding to the text information; using the voice voiceprint information to test the voiceprint recognition system. Through the present application, the problem of low efficiency of voiceprint recognition test in the related art is solved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and more specifically, to a voiceprint recognition testing method, device, storage medium and electronic device. Background Art

[0002] In the field of biometrics, voiceprint recognition technology has gradually become an important and popular authentication method in people's daily life and work due to its unique convenience, economy and accuracy. The security test of voiceprint recognition has also become a research hotspot. Among them, voiceprint recording playback attack is a typical test method, which records the user audio in advance and plays the audio during the voiceprint recognition process for testing. However, the common voiceprint recognition currently often uses random content voiceprint technology. Its principle is to prompt the user to read aloud the randomly displayed numbers or text during the voiceprint recognition process. If the content read aloud is inconsistent, the recognition fails. This method makes the pre-recorded audio invalid. Therefore, during the test process, the tester needs to manually identify the content to be read aloud and synthesize the recording in a short time. The manual operation is cumbersome and the testing cost is high.

[0003] With regard to the problem of low efficiency of voiceprint recognition testing in related technologies, no effective solution has been proposed so far. Summary of the invention

[0004] The main purpose of the present application is to provide a voiceprint recognition test method, device, storage medium and electronic device to solve the problem of low efficiency of voiceprint recognition test in related technologies.

[0005] In order to achieve the above-mentioned purpose, according to one aspect of the present application, a voiceprint recognition test method is provided. The method comprises: obtaining an image frame in a target video stream, wherein the image frame includes text information; extracting the text information in the image frame; using the text information to match in a preset voice library to obtain the voice voiceprint information corresponding to the text information; and using the voice voiceprint information to test the voiceprint recognition system.

[0006] Furthermore, before obtaining the image frames in the target video stream, the method also includes: obtaining an initial video stream through a target device; encoding the initial video stream to obtain a processed video stream; and transmitting the processed video stream through a socket.

[0007] Furthermore, before extracting the text information in the image frame, the method also includes: processing the text information in the image frame according to a preset processing method to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing and character segmentation processing.

[0008] Furthermore, extracting text information from image frames includes: performing feature extraction on the processed image frames through a convolutional neural network to obtain feature information of the image frames; inputting the feature information and weight information of the target model into a convolutional recurrent neural network and performing encoding processing to obtain the probability of a character set in the text information; and determining the text information in the image frame based on the probability of the character set.

[0009] Furthermore, before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and performing encoding processing, the method also includes: obtaining the state information of the neurons of the recurrent neural network and the output result of the target model at the previous moment; determining the weight information of the target model based on the state information of the neurons and the output result.

[0010] Furthermore, using text information to match in a preset voice library to obtain voice print information corresponding to the text information includes: performing noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; segmenting the processed voice audio to obtain segmented audio information; judging whether there are digital or text audio segments in the segmented audio information within a preset time; if there are digital or text audio segments in the segmented audio information, matching the digital or text audio segments with the target voice print content to obtain voice print information corresponding to the text information.

[0011] Furthermore, after testing the voiceprint recognition system using speech voiceprint information, the method also includes: obtaining the test results, obtaining a result display picture from the test results; verifying the result display picture through a target detection algorithm to obtain a verified picture; determining the test result of voiceprint recognition based on the verified picture, wherein the test result of voiceprint recognition includes at least a successful test of voiceprint recognition and a failed test of voiceprint recognition.

[0012] In order to achieve the above-mentioned purpose, according to another aspect of the present application, a test device for voiceprint recognition is provided. The device comprises: a first acquisition unit, used to acquire an image frame in a target video stream, wherein the image frame includes text information; a first extraction unit, used to extract text information in the image frame; a first matching unit, used to use the text information to match in a preset voice library to obtain voice voiceprint information corresponding to the text information; and a first testing unit, used to use the voice voiceprint information to test the voiceprint recognition system.

[0013] Furthermore, the device includes: a second acquisition unit, used to acquire an initial video stream through a target device before acquiring image frames in a target video stream; a first processing unit, used to encode the initial video stream to obtain a processed video stream; and a first transmission unit, used to transmit the processed video stream through a socket.

[0014] Furthermore, the device also includes: a second processing unit, which is used to process the text information in the image frame according to a preset processing method before extracting the text information in the image frame to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing and character segmentation processing.

[0015] Furthermore, the first extraction unit includes: a first extraction module, used to perform feature extraction on the processed image frame through a convolutional neural network to obtain feature information of the image frame; a first processing module, used to input the feature information and weight information of the target model into the convolutional recurrent neural network and perform encoding processing to obtain the probability of the character set in the text information; a first determination module, used to determine the text information in the image frame according to the probability of the character set.

[0016] Furthermore, the device also includes: a third acquisition unit, used to obtain the state information of the neurons of the recurrent neural network and the output result of the target model at the previous moment before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and encoding processing; a first determination unit, used to determine the weight information of the target model based on the state information of the neurons and the output result.

[0017] Furthermore, the first matching unit includes: a second processing module, which is used to perform noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; a third processing module, which is used to segment the processed voice audio to obtain segmented audio information; a first judgment module, which is used to judge whether there are digital or text audio fragments in the segmented audio information within a preset time; the first matching module is used to match the digital or text audio fragment with the target voiceprint content if there are digital or text audio fragments in the segmented audio information, so as to obtain the voice voiceprint information corresponding to the text information.

[0018] Furthermore, the device also includes: a fourth acquisition unit, used to obtain the test result after testing the voiceprint recognition system using voice voiceprint information, and obtain a result display picture from the test result; a first verification unit, used to verify the result display picture through a target detection algorithm to obtain a verified picture; a second determination unit, used to determine the test result of voiceprint recognition based on the verified picture, wherein the test result of voiceprint recognition at least includes a successful test of voiceprint recognition and a failed test of voiceprint recognition.

[0019] According to another aspect of an embodiment of the present application, an electronic device is also provided, including one or more processors and a memory, and the computer program / instructions execute any of the above methods when executed by the processor.

[0020] According to another aspect of an embodiment of the present application, a computer-readable storage medium is provided, on which a computer program / instruction is stored. When the computer program / instruction is executed by a processor, any one of the above methods is performed.

[0021] Through this application, the following steps are adopted: obtaining an image frame in a target video stream, wherein the image frame includes text information; extracting the text information in the image frame; using the text information to match in a preset voice library to obtain the voice voiceprint information corresponding to the text information; using the voice voiceprint information to test the voiceprint recognition system. The problem of low efficiency of voiceprint recognition testing in the related art is solved, by using text information to match in a preset voice library, and then automatically generating the voice voiceprint information corresponding to the text information, thereby achieving the effect of improving the efficiency of voiceprint recognition testing. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 is a flow chart of a voiceprint recognition testing method provided according to an embodiment of the present application;

[0024] Figure 2 It is a system workflow diagram of the voiceprint recognition testing method provided in an embodiment of the present application;

[0025] Figure 3 Schematic diagram of a test device for voiceprint recognition according to an embodiment of the present application Figure 1 ;

[0026] Figure 4 Schematic diagram of a test device for voiceprint recognition according to an embodiment of the present application Figure 2 ;

[0027] Figure 5 is a schematic diagram of text recognition of a voiceprint recognition test device provided in an embodiment of the present application;

[0028] Figure 6 It is a schematic diagram of audio splicing of a test device for voiceprint recognition provided in an embodiment of the present application;

[0029] Figure 7 It is a schematic diagram of comparing the results of the voiceprint recognition test device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0030] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0031] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] For the convenience of description, some nouns or terms involved in the embodiments of the present application are explained below:

[0034] ADB (Android Debug Bridge): is a device communication technology that can perform various mobile device operations through the PC.

[0035] Socket: refers to a socket, which is an abstraction of an endpoint for two-way communication between application processes on different hosts in a network.

[0036] FFMPEG: is a set of open source computer programs that can be used to record, convert digital audio and video, and convert them into streams.

[0037] According to an embodiment of the present application, a test method for voiceprint recognition is provided.

[0038] Figure 1 FIG. 1 is a flow chart of a method for testing voiceprint recognition according to an embodiment of the present application. Figure 1 As shown, the method comprises the following steps:

[0039] Step S101, obtaining an image frame in a target video stream, wherein the image frame includes text information.

[0040] Specifically, the image frames in the target video stream are acquired in real time. Before the acquisition, the target video stream needs to be processed.

[0041] Optionally, in the test method for voiceprint recognition provided in an embodiment of the present application, before obtaining the image frame in the target video stream, the method also includes: obtaining an initial video stream through the target device; encoding the initial video stream to obtain a processed video stream; and transmitting the processed video stream through a socket.

[0042] Specifically, the initial screen video stream is read through the MediaCodec class and VirtualDisplay class provided by the Android system of the mobile phone (corresponding to the target device in this application), and then the H264 encoding processing operation is performed on the initial screen video stream to obtain the processed video stream (corresponding to the target video stream in this application), and then the instructions of the command line are executed by ADB and transmitted to the computer side in real time through Socket. The computer side receives the video stream transmitted by the mobile phone through Socket, and uses FFMPEG multimedia framework to decode and display it. By acquiring the data in the video stream in real time, the efficiency of subsequent extraction of text information is further improved.

[0043] Step S102: extracting text information from the image frame.

[0044] Specifically, the text information of the image frame of the present application includes the text information to be read aloud, and the text information to be read aloud is identified and extracted. Before extracting the text information to be read aloud, the text information in the image frame needs to be processed.

[0045] Optionally, in the test method for voiceprint recognition provided in an embodiment of the present application, before extracting the text information in the image frame, the method also includes: processing the text information in the image frame according to a preset processing method to obtain a processed image frame, wherein the preset processing method includes at least: image binarization processing, image noise reduction processing and character segmentation processing.

[0046] Specifically, the present application obtains a processed image frame by binarizing, denoising and character segmenting a screenshot containing the text to be read aloud (corresponding to the image frame in the present application), thereby reducing useless information in the image and improving the recognition efficiency of text information.

[0047] Optionally, in the test method for voiceprint recognition provided in an embodiment of the present application, extracting text information from an image frame includes: performing feature extraction on the processed image frame through a convolutional neural network to obtain feature information of the image frame; inputting the feature information and weight information of the target model into a convolutional recurrent neural network and performing encoding processing to obtain the probability of a character set in the text information; and determining the text information in the image frame based on the probability of the character set.

[0048] Specifically, the convolutional neural network automatically extracts feature information from the input processed image frame, combines the feature information with the weight information of the attention model (corresponding to the target model in this application), inputs the recurrent neural network for encoding, obtains the probability distribution of the character set of the entire text information, and extracts the character with the highest probability in the character set as the recognition result of the text information, and extracts the recognition result. This application recognizes text information by obtaining the probability of the character set, making the recognition result of the text information more accurate.

[0049] Optionally, in the test method for voiceprint recognition provided in an embodiment of the present application, before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and performing encoding processing, the method also includes: obtaining the state information of the neurons of the recurrent neural network and the output result of the target model at the previous moment; determining the weight information of the target model based on the state information of the neurons and the output result.

[0050] Specifically, the attention model calculates the attention weight according to the hidden state of the neurons in the recurrent neural network and the output result of the attention model at the previous moment, obtains the calculation result, and uses the calculation result as the weight information of the attention model, which further reflects the accuracy of the present application in recognizing text information.

[0051] Step S103, matching the text information in a preset voice database to obtain voice print information corresponding to the text information.

[0052] Specifically, the extracted text information is concatenated and matched with the audio in the recording corpus to obtain the speech voiceprint information corresponding to the text information, and then the speech voiceprint information is played as the recorded audio.

[0053] Optionally, in the test method for voiceprint recognition provided in the embodiment of the present application, text information is matched in a preset voice library to obtain the voice voiceprint information corresponding to the text information, including: performing noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; segmenting the processed voice audio to obtain segmented audio information; judging whether there are digital or text audio fragments in the segmented audio information within a preset time; if there are digital or text audio fragments in the segmented audio information, matching the digital or text audio fragment with the target voiceprint content to obtain the voice voiceprint information corresponding to the text information.

[0054] Specifically, the speech audio in the preset speech library is automatically de-noised by using the spectral subtraction method, where it is assumed that r(t), x(t), and n(t) are the sound signal after denoising, the sound signal before denoising, and the background noise signal, respectively. The relationship in the time domain is:

[0055] r(t)=x(t)-n(t)

[0056] Perform Fourier transform on the three functions respectively, and the corresponding frequency domain signals are R(w), X(w), and N(w). Then the relationship in the frequency domain is:

[0057] R(w)=X(w)-N(w)

[0058] According to the target text content, the noise-reduced speech audio is segmented and reorganized to generate the voiceprint audio required for the current device login, wherein audio segmentation mainly processes the long audio in the corpus, and judges whether the time window is a digital or text interval based on the average peak size of the time domain waveform in a certain time window. If there are numbers or text audio segments in the segmented audio information, the audio is segmented in the window, and the recording corpus is updated according to the recording corpus label, wherein audio reorganization extracts single-digit or text audio segments in the corpus according to the target voiceprint content and splices them together for reorganization. This application automatically realizes the segmentation and reorganization of audio to obtain the voice voiceprint information to be tested, further improving the efficiency of voice voiceprint recognition and reducing testing costs.

[0059] Step S104: testing the voiceprint recognition system using the speech voiceprint information.

[0060] Specifically, the voiceprint recognition system is tested using the voiceprint information to be tested, and a sign image of success or failure of the test feedback is extracted according to the target detection, so as to determine the result of this automated test.

[0061] Optionally, in the test method for voiceprint recognition provided in an embodiment of the present application, after the voiceprint recognition system is tested using speech voiceprint information, the method further includes: obtaining the test results, obtaining a result display picture from the test results; verifying the result display picture through a target detection algorithm to obtain a verified picture; determining the test result of voiceprint recognition based on the verified picture, wherein the test result of voiceprint recognition includes at least a successful test of voiceprint recognition and a failed test of voiceprint recognition.

[0062] Specifically, the present application adopts the YOLO target detection neural network function in the real-time target detection algorithm to identify the return result of the voiceprint information verification (corresponding to the result display picture in the present application). Specifically, the input of the target detection algorithm includes the screen display image of the voiceprint verification result, and the output recognition result (corresponding to the test result in the present application) includes the position coordinates, width, height and category of the successful or failed logo pattern. By analyzing the logo pattern position information and category information in the test result, it is further determined whether the voiceprint recognition test is successful or the voiceprint recognition test fails.

[0063] Specifically, the above-mentioned system operation time information, target voiceprint content, synthesized voiceprint content, return value and other attributes are stored in a structured manner, and the results are statistically analyzed after the completion of the automated test, and the statistical results are displayed intuitively, thereby improving the ability of automated test implementers to clearly understand the results of the voiceprint test.

[0064] Optionally, Figure 2 is a system workflow diagram of the voiceprint recognition testing method provided in the embodiment of the present application, such as Figure 2 As shown, the mobile phone screen is captured in real time and the image is transmitted to the computer. The OCR text recognition model is used to recognize the text to be read displayed on the screen. The audio splicing module is used to synthesize the target voiceprint and play it. The result is captured and detected to obtain the test result, which is saved before the next automated test process is carried out.

[0065] In summary, the test method for voiceprint recognition provided by the embodiment of the present application obtains image frames in the target video stream, wherein the image frames include text information; extracts the text information in the image frames; uses the text information to match in a preset voice library to obtain the voice and voiceprint information corresponding to the text information; and uses the voice and voiceprint information to test the voiceprint recognition system. The problem of low efficiency of voiceprint recognition testing in the related art is solved, and the voice and voiceprint information corresponding to the text information is automatically generated by using the text information to match in a preset voice library, thereby achieving the effect of improving the efficiency of voiceprint recognition testing.

[0066] It should be noted that the steps shown in the flowcharts of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and that, although a logical order is shown in the flowcharts, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0067] The embodiment of the present application also provides a test device for voiceprint recognition. It should be noted that the test device for voiceprint recognition in the embodiment of the present application can be used to execute the test method for voiceprint recognition provided in the embodiment of the present application. The test device for voiceprint recognition provided in the embodiment of the present application is introduced below.

[0068] Figure 3 Schematic diagram of a test device for voiceprint recognition according to an embodiment of the present application Figure 1 .like Figure 3 As shown, the device includes: a first acquisition unit 301, a first extraction unit 302, a first matching unit 303, and a first testing unit 304.

[0069] Specifically, the first acquisition unit 301 is used to acquire an image frame in a target video stream, wherein the image frame includes text information;

[0070] A first extraction unit 302, used to extract text information from the image frame;

[0071] The first matching unit 303 is used to match the text information in a preset voice database to obtain the voice print information corresponding to the text information;

[0072] The first testing unit 304 is used to test the voiceprint recognition system using the speech voiceprint information.

[0073] In summary, the voiceprint recognition test device provided in the embodiment of the present application is used to obtain image frames in the target video stream through the first acquisition unit 301, wherein the image frames include text information; the first extraction unit 302 is used to extract text information in the image frames; the first matching unit 303 is used to use the text information to match in a preset voice library to obtain the voice voiceprint information corresponding to the text information; the first testing unit 304 is used to test the voiceprint recognition system using the voice voiceprint information, which solves the problem of low efficiency of voiceprint recognition testing in related technologies, and uses the text information to match in a preset voice library, and then automatically generates the voice voiceprint information corresponding to the text information, thereby achieving the effect of improving the efficiency of voiceprint recognition testing.

[0074] Optionally, in the test device for voiceprint recognition provided in an embodiment of the present application, the device includes: a second acquisition unit, used to acquire an initial video stream through a target device before acquiring image frames in a target video stream; a first processing unit, used to encode the initial video stream to obtain a processed video stream; and a first transmission unit, used to transmit the processed video stream through a socket.

[0075] Optionally, in the voiceprint recognition test device provided in an embodiment of the present application, the device also includes: a second processing unit, used to process the text information in the image frame according to a preset processing method before extracting the text information in the image frame to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing and character segmentation processing.

[0076] Optionally, in the voiceprint recognition test device provided in the embodiment of the present application, the first extraction unit 302 includes: a first extraction module, used to perform feature extraction on the processed image frame through a convolutional neural network to obtain feature information of the image frame; a first processing module, used to input the feature information and weight information of the target model into the convolutional recurrent neural network and perform encoding processing to obtain the probability of the character set in the text information; a first determination module, used to determine the text information in the image frame according to the probability of the character set.

[0077] Optionally, in the test device for voiceprint recognition provided in an embodiment of the present application, the device also includes: a third acquisition unit, used to obtain the state information of the neurons of the recurrent neural network and the output result of the target model at the previous moment before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and encoding processing; a first determination unit, used to determine the weight information of the target model based on the state information of the neurons and the output result.

[0078] Optionally, in the voiceprint recognition test device provided in the embodiment of the present application, the first matching unit 303 includes: a second processing module, which is used to perform noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; a third processing module, which is used to segment the processed voice audio to obtain segmented audio information; a first judgment module, which is used to judge whether there are digital or text audio fragments in the segmented audio information within a preset time; and a first matching module, which is used to match the digital or text audio fragment with the target voiceprint content if there are numbers or text audio fragments in the segmented audio information, so as to obtain the voice voiceprint information corresponding to the text information.

[0079] Optionally, in the voiceprint recognition test device provided in the embodiment of the present application, the device also includes: a fourth acquisition unit, used to obtain the test result after testing the voiceprint recognition system using voice voiceprint information, and obtain a result display picture from the test result; a first verification unit, used to verify the result display picture through a target detection algorithm to obtain a verified picture; a second determination unit, used to determine the test result of voiceprint recognition based on the verified picture, wherein the test result of voiceprint recognition includes at least a success in the voiceprint recognition test and a failure in the voiceprint recognition test.

[0080] Optionally, Figure 4 Schematic diagram of a test device for voiceprint recognition according to an embodiment of the present application Figure 2 ,like Figure 4 As shown, the device consists of three parts: text recognition module, audio splicing module, and result comparison module.

[0081] Optionally, Figure 5 is a schematic diagram of text recognition of a test device for voiceprint recognition according to an embodiment of the present application, such as Figure 5 As shown, the text recognition module mainly consists of two parts: a screen capture unit and an OCR recognition unit. Specifically, the text recognition module is responsible for capturing the mobile phone screen in real time and transmitting it to the computer. The computer recognizes the text to be read displayed on the screen and transmits it to the audio splicing module for processing.

[0082] Optionally, Figure 6 is a schematic diagram of audio splicing of a test device for voiceprint recognition provided in an embodiment of the present application, such as Figure 6 As shown, the audio splicing module mainly consists of two parts: a background noise removal unit and an audio segmentation and reorganization unit. Specifically, the audio splicing module is responsible for splicing the audio in the recording corpus according to the text information obtained by the text recognition module, generating a recording playback to test the voiceprint recognition, and passing the test results to the result comparison module.

[0083] Optionally, Figure 7 is a schematic diagram of comparing the results of the voiceprint recognition test device provided in the embodiment of the present application, such as Figure 7 As shown, the result comparison module includes an AI recognition unit and a result storage unit. Specifically, the result comparison module is used to capture the voiceprint test result image, extract keywords and key graphics to determine whether the voiceprint verification is successful, and finally save the judgment result.

[0084] The test device for voiceprint recognition includes a processor and a memory. The first acquisition unit 301, the first extraction unit 302, the first matching unit 303, the first test unit 304, etc. are all stored in the memory as program units, and the processor executes the program units stored in the memory to implement corresponding functions.

[0085] The processor contains a kernel, which calls the corresponding program unit from the memory. One or more kernels can be set, and the voiceprint recognition test can be performed by adjusting the kernel parameters.

[0086] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one memory chip.

[0087] An embodiment of the present invention provides a storage medium on which a program is stored. When the program is executed by a processor, a test method for voiceprint recognition is implemented.

[0088] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes a test method for voiceprint recognition when the program is running.

[0089] An embodiment of the present invention provides a device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining image frames in a target video stream, wherein the image frames include text information; extracting the text information in the image frames; using the text information to match in a preset voice library to obtain voice voiceprint information corresponding to the text information; and using the voice voiceprint information to test a voiceprint recognition system.

[0090] When executing the program, the processor also implements the following steps: before obtaining the image frame in the target video stream, obtaining the initial video stream through the target device; encoding the initial video stream to obtain a processed video stream; and transmitting the processed video stream through the socket.

[0091] When the processor executes the program, the following steps are also implemented: before extracting the text information in the image frame, the text information in the image frame is processed according to a preset processing method to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing and character segmentation processing.

[0092] When the processor executes the program, the following steps are also implemented: feature extraction is performed on the processed image frame through a convolutional neural network to obtain feature information of the image frame; the feature information and weight information of the target model are input into the convolutional recurrent neural network and encoded to obtain the probability of the character set in the text information; the text information in the image frame is determined according to the probability of the character set.

[0093] When the processor executes the program, the following steps are also implemented: before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and encoding processing, the state information of the neurons of the recurrent neural network and the output result of the target model at the previous moment are obtained; the weight information of the target model is determined based on the state information of the neurons and the output result.

[0094] When the processor executes the program, the following steps are also implemented: performing noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; segmenting the processed voice audio to obtain segmented audio information; judging whether there are digital or text audio segments in the segmented audio information within a preset time; if there are digital or text audio segments in the segmented audio information, matching the digital or text audio segments with the target voiceprint content to obtain the voice voiceprint information corresponding to the text information.

[0095] When the processor executes the program, the following steps are also implemented: after testing the voiceprint recognition system using voice voiceprint information, the test results are obtained, and a result display image is obtained from the test results; the result display image is verified through a target detection algorithm to obtain a verified image; and the test result of the voiceprint recognition is determined based on the verified image, wherein the test result of the voiceprint recognition includes at least a successful test of the voiceprint recognition and a failed test of the voiceprint recognition.

[0096] The devices in this article can be servers, PCs, PADs, mobile phones, etc.

[0097] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: obtaining image frames in a target video stream, wherein the image frames include text information; extracting the text information in the image frames; using the text information to match in a preset voice library to obtain voice voiceprint information corresponding to the text information; and using the voice voiceprint information to test the voiceprint recognition system.

[0098] When executed on a data processing device, it is also suitable for executing an initialization program having the following method steps: before obtaining the image frame in the target video stream, obtaining the initial video stream through the target device; encoding the initial video stream to obtain a processed video stream; and transmitting the processed video stream through a socket.

[0099] When executed on a data processing device, it is also suitable for executing an initialized program having the following method steps: before extracting text information in an image frame, processing the text information in the image frame according to a preset processing method to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing, and character segmentation processing.

[0100] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: extracting features of the processed image frame through a convolutional neural network to obtain feature information of the image frame; inputting the feature information and weight information of the target model into the convolutional recurrent neural network and performing encoding processing to obtain the probability of the character set in the text information; and determining the text information in the image frame based on the probability of the character set.

[0101] When executed on a data processing device, it is also suitable for executing an initialization program having the following method steps: before inputting feature information and weight information of a target model into a target convolutional recurrent neural network and performing encoding processing, obtaining the state information of neurons of the recurrent neural network and the output result of the target model at the previous moment; determining the weight information of the target model based on the state information of the neurons and the output result.

[0102] When executed on a data processing device, it is also suitable for executing a program that is initialized with the following method steps: performing noise reduction processing on the voice audio in a preset voice library to obtain processed voice audio; segmenting the processed voice audio to obtain segmented audio information; judging whether there are digital or text audio fragments in the segmented audio information within a preset time; if there are digital or text audio fragments in the segmented audio information, matching the digital or text audio fragment with the target voiceprint content to obtain the voice voiceprint information corresponding to the text information.

[0103] When executed on a data processing device, it is also suitable for executing an initialized program having the following method steps: after testing the voiceprint recognition system using speech voiceprint information, obtaining the test results, and obtaining a result display picture from the test results; verifying the result display picture through a target detection algorithm to obtain a verified picture; determining the test result of voiceprint recognition based on the verified picture, wherein the test result of voiceprint recognition includes at least a successful test of voiceprint recognition and a failed test of voiceprint recognition.

[0104] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0105] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0106] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0107] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0108] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0109] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0110] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0111] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0112] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0113] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A voiceprint recognition testing method, characterized in that: include: Acquire an image frame in a target video stream, wherein the image frame includes text information; Extracting text information from the image frame; The text information is matched in a preset voice database to obtain voice print information corresponding to the text information; Using the voiceprint information to test the voiceprint recognition system; The text information is matched in a preset voice database to obtain voice print information corresponding to the text information, including: Performing noise reduction processing on the speech audio in the preset speech library to obtain processed speech audio; Segmenting the processed voice audio to obtain segmented audio information; Determining whether there is a digital or text audio segment in the segmented audio information within a preset time; If the number or the text audio segment exists in the segmented audio information, the number or the text audio segment is matched with the target voiceprint content to obtain the voice voiceprint information corresponding to the text information.

2. The method according to claim 1, characterized in that Before acquiring the image frame in the target video stream, the method further includes: Get the initial video stream from the target device; Encoding the initial video stream to obtain a processed video stream; The processed video stream is transmitted through a socket.

3. The method according to claim 1, characterized in that Before extracting the text information in the image frame, the method further includes: The text information in the image frame is processed according to a preset processing method to obtain a processed image frame, wherein the preset processing method at least includes: image binarization processing, image noise reduction processing and character segmentation processing.

4. The method according to claim 3, characterized in that Extracting text information from the image frame includes: Extracting features from the processed image frame by a convolutional neural network to obtain feature information of the image frame; Inputting the feature information and the weight information of the target model into a convolutional recurrent neural network and performing encoding processing to obtain the probability of the character set in the text information; The text information in the image frame is determined according to the probability of the character set.

5. The method according to claim 4, characterized in that Before inputting the feature information and the weight information of the target model into the target convolutional recurrent neural network and performing encoding processing, the method further includes: Obtaining state information of neurons of the recurrent neural network and an output result of the target model at a previous moment; The weight information of the target model is determined according to the state information of the neuron and the output result.

6. The method according to claim 1, characterized in that After the voiceprint recognition system is tested using the speech voiceprint information, the method further includes: Obtaining test results, and obtaining result display pictures from the test results; Verifying the result display image through a target detection algorithm to obtain a verified image; The test result of the voiceprint recognition is determined according to the verified image, wherein the test result of the voiceprint recognition at least includes a success of the voiceprint recognition test and a failure of the voiceprint recognition test.

7. A test device for voiceprint recognition, characterized in that: include: A first acquisition unit, configured to acquire an image frame in a target video stream, wherein the image frame includes text information; A first extraction unit, used to extract text information from the image frame; A first matching unit, configured to match the text information in a preset voice database to obtain voice print information corresponding to the text information; A first testing unit, used to test the voiceprint recognition system using the voice voiceprint information; Among them, the first matching unit includes: a second processing module, which is used to perform noise reduction processing on the voice audio in the preset voice library to obtain processed voice audio; a third processing module, which is used to segment the processed voice audio to obtain segmented audio information; a first judgment module, which is used to judge whether there are digital or text audio fragments in the segmented audio information within a preset time; and a first matching module, which is used to match the digital or text audio fragment with the target voiceprint content if the digital or text audio fragment exists in the segmented audio information to obtain the voice voiceprint information corresponding to the text information.

8. A computer-readable storage medium, characterized in that: The storage medium includes a stored program, wherein the program executes the method according to any one of claims 1 to 6.

9. An electronic device, characterized in that: include: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the voiceprint recognition test method described in any one of claims 1 to 6.

10. A computer program product, characterized in that A program adapted to execute the steps of initiating the method as claimed in any one of claims 1 to 6 when executed on a data processing device.

Images