A method for identifying authenticity of video content

Abstract

The present invention discloses a method for identifying the authenticity of video content, which includes the following steps: performing video preprocessing on a video to be identified, and obtaining video content identification data frame by frame, wherein the video content identification data includes human skin color data and environmental impact data; establishing a human skin color comparison model, substituting the obtained human skin color data into the human skin color comparison model for human skin color analysis, so as to obtain a human skin color change mark; establishing an environmental data comparison model, substituting the obtained human skin color data and environmental impact data into the environmental data comparison model for environmental impact analysis, and generating an environmental impact coefficient and a human skin color coefficient; establishing a deep learning model, analyzing and comparing the environmental impact coefficient and the human skin color impact coefficient, so as to obtain an image processing mark; and generating a false content mark, a suspicious content mark, and a credible content mark based on the human skin color change mark and the image processing mark.

Description

Technical Field

[0001] The present invention relates to the technical field of video authenticity identification, and in particular to a method for identifying the authenticity of video content. Background Art

[0002] With the rapid development of self-media today and the gradual popularization of various video software on the Internet, the number of people using short videos has become increasingly wide.

[0003] However, the quality of short videos nowadays varies greatly. Ordinary users are unable to effectively analyze videos and are easily misled by various types of false videos. In addition, some criminals create false evidence by synthesizing false videos, which increases the difficulty for public officials to handle cases. Therefore, it can be seen that in-depth research on a method for authenticating video content is necessary.

[0004] After searching, the video authenticity inspection and identification method and system based on time watermark change analysis proposed by publication number CN113034430A adopted the identification method 1 of comparing videos to parse the pictures corresponding to different frames, and then analyzing the digital content of the pictures corresponding to different frames. The second method analyzes the changes of the highlights of each frame in the video on the basis of method 1, and achieves the authenticity identification of the video by re-analyzing the video.

[0005] In light of the above, the following issues remain:

[0006] 1. Most existing video authentication methods analyze multiple video frames and then use computer vision and image processing technology to compare and analyze the brightness, texture, color tone, and metadata (shooting date, time, and location) of the image corresponding to each frame to obtain the video authenticity verification result. This has the limitation of single video authentication method.

[0007] 2. In addition, during the process of video authenticity verification, especially when verifying the authenticity of videos containing portraits, the creator of the video may edit each frame one by one without modifying the background data. In this case, the method of verifying the authenticity of the video based on brightness, texture, and hue will have certain loopholes, resulting in accidental errors in the authenticity verification.

[0008] In order to solve the above-mentioned problems, a method for identifying the authenticity of video content is proposed. Summary of the Invention

[0009] The purpose of the present invention is to provide a method for identifying the authenticity of video content to solve the shortcomings of the background technology.

[0010] In order to achieve the above object, the present invention provides the following technical solutions:

[0011] The method for identifying the authenticity of video content comprises the following steps:

[0012] Step S100: Preprocessing the video to be identified and obtaining video content identification data frame by frame. The video content identification data includes human skin color data and environmental impact data.

[0013] Step S200: establishing a human skin color comparison model, substituting the acquired human skin color data into the human skin color comparison model to perform human skin color analysis to obtain a human skin color change mark;

[0014] Step S300: establishing an environmental data comparison model, substituting the acquired human skin color data and environmental impact data into the environmental data comparison model to perform environmental impact analysis, and generating an environmental impact coefficient and a human skin color coefficient;

[0015] Step S400: establishing a deep learning model, analyzing and comparing the environmental impact coefficient and the human skin color impact coefficient to obtain an image processing mark;

[0016] Step S500: Analyze and generate false content identification, suspicious content identification and credible content identification based on the human skin color change mark and the image processing mark, and mark the matching identification interval on the video editing software.

[0017] In a preferred embodiment, the human skin color data includes facial skin color data X and body skin color data Y, and the environmental impact data includes an ambient light intensity factor L, an ambient reflection factor R, and a light source angle factor F.

[0018] In a preferred embodiment, the process of analyzing and generating a human skin color change mark is as follows:

[0019] The video to be identified is divided into n identification intervals, where n is an integer greater than 1. A single identification interval consists of two frames of images. Images of frames k and k+1 within the n identification intervals are obtained respectively. Facial skin color data comparison threshold reference values ​​X1 and X2 are set, where X1<X2 and both X1 and X2 are greater than 0.

[0020] Get the facial skin color data X at frame k k and the facial skin color data X at frame k+1 k+1 , generate the facial skin color data difference Xc between frame k and frame k+1 through data processing, where X=X k+1 -X k (X k 、X k+1is greater than 0, Xc is greater than or equal to 0), substituting the facial skin color data difference Xc into the facial skin color comparison threshold for comparative analysis to generate facial skin color level information, the facial skin color level information including primary facial skin color information, secondary facial skin color information and tertiary facial skin color information;

[0021] Set the body skin color data comparison threshold reference values ​​Y1 and Y2, where Y1 < Y2, and both Y1 and Y2 are greater than 0;

[0022] Get the body skin color data Y at frame k k and the body skin color data Y at frame k+1 k+1 , generate the difference Yc of the body skin color data between frame k and frame k+1 through data processing, where Y=Y k+1 -Y k (Y k 、Y k+1 is greater than 0, Y is greater than or equal to 0), the body skin color data difference Yc is substituted into the body skin color comparison threshold for comparative analysis to generate body skin color grade information, which includes primary body skin color information, secondary body skin color information and tertiary body skin color information.

[0023] In a preferred embodiment, the facial skin color level information generation logic is:

[0024] When the facial skin color data difference value Xc is less than the threshold reference value X1, the identification interval where the facial skin color data difference value Xc is located is marked as the first-level facial skin color information; when the facial skin color data difference value Xc is greater than the threshold reference value X1 and the facial skin color data difference value Xc is less than the threshold reference value X2, the identification interval where the facial skin color data difference value Xc is located is marked as the second-level facial skin color information; when the facial skin color data Xc is greater than the threshold reference value X2, the identification interval where the facial skin color data difference value Xc is located is marked as the third-level facial skin color information;

[0025] The logic for generating the body skin color level information is as follows:

[0026] When the body skin color data difference Yc is less than the threshold reference value Y1, the identification interval where the body skin color data difference Yc is located is marked as the first-level body skin color information; when the body skin color data is greater than the threshold reference value Y1 and the body skin color data difference Yc is less than the threshold reference value Y2, the identification interval where the body skin color data difference Yc is located is marked as the second-level body skin color information; when the body skin color data difference Yc is greater than the threshold reference value Y2, the identification interval where the body skin color data difference Yc is located is marked as the third-level body skin color information.

[0027] In a preferred embodiment, the generation logic of the environmental impact coefficient is:

[0028] A combined model of illumination error coefficients is designed in the environmental data comparison model, weights are assigned to each molecule, and the environmental impact coefficient is calculated based on formulaic analysis.

[0029] In a preferred embodiment, the logic for generating the human skin color coefficient is:

[0030] The human skin color change mark includes a primary skin color change mark, an intermediate skin color change mark, and an advanced skin color change mark. A skin color distribution model is set in the environmental data comparison model. The facial skin color data X and the body skin color data Y are analyzed and processed in a formulaic manner to generate a human skin color coefficient βn. The root processing is performed by taking the ratio of the absolute value of the difference between the facial skin color data X and Y and Y and adding the skin color error ratio revision constant K.

[0031] In a preferred embodiment, the logic for generating the image processing mark is:

[0032] The image processing mark includes a primary image processing mark, a secondary image processing mark, and a tertiary image processing mark. The environmental impact coefficient and the human skin color impact coefficient are analyzed and compared by formulating, the influence of the environment in the image is removed, and the original image coefficient γn is generated. The original image coefficient comparison thresholds W1 and W2 are set, where W1>W2>0. The original image coefficient γn is substituted into the original image coefficient comparison threshold for analysis:

[0033] When the original image coefficient γn of a single identification area is less than the original image coefficient comparison threshold W2, a first-level image processing mark is generated for the identification area. When the original image coefficient γn of a single identification area is greater than the original image coefficient comparison threshold W2 and less than the original image coefficient comparison threshold W1, a second-level image processing mark is generated for the identification area. When the original image coefficient γn of a single identification area is greater than the original image coefficient comparison threshold W1, a third-level image processing mark is generated for the identification area.

[0034] In a preferred embodiment, the false content identifier, suspicious content identifier, and credible content identifier are generated based on the analysis of the human skin color change mark and the image processing mark. The generation logic is as follows:

[0035] When a single identification area contains both a primary image processing mark and a mid-level skin color change mark, a primary image processing mark and a high-level skin color change mark, or a tertiary image processing mark and a primary skin color change mark, a false content mark is generated for the identification area;

[0036] When a secondary image processing mark and a primary skin color change mark, a secondary image processing mark and a high-level skin color change mark, a tertiary image processing mark and a mid-level skin color change mark, or a tertiary environmental impact mark and a high-level skin color change mark coexist within a single identification interval, a suspicious content mark is generated for the identification area;

[0037] When a single identification interval contains a primary image processing mark and a primary skin color change mark, a secondary image processing mark and a mid-level skin color change mark, or a tertiary image processing mark and a high-level skin color change mark, a trusted content identifier is generated for the identification area.

[0038] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0039] The above-mentioned implementation analyzes the authentication video frame by frame, utilizing the skin color difference ratio between facial and body skin tones to achieve more accurate video authenticity verification for retouched portraits. Furthermore, secondary processing of the primary authentication result, analyzing and processing environmental factors with the primary authentication result, further enhances the accuracy of authentication results and the diversity of authentication methods in the authentication of video content containing portraits. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, a brief introduction to the drawings required for use in the embodiments will be given below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0041] Figure 1 The present invention is a flowchart of a method for identifying the authenticity of video content. DETAILED DESCRIPTION

[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0043] See also Figure 1 As shown, this embodiment provides a method for identifying the authenticity of video content, more specifically, a method for identifying the authenticity of video content with a human portrait. The method includes the following steps:

[0044] Step S100: Preprocessing the video to be identified and obtaining video content identification data frame by frame. The video content identification data includes human skin color data and environmental impact data.

[0045] Step S200: establishing a human skin color comparison model, substituting the acquired human skin color data into the human skin color comparison model to perform human skin color analysis to obtain a human skin color change mark;

[0046] Step S300: establishing an environmental data comparison model, substituting the acquired human skin color data and environmental impact data into the environmental data comparison model to perform environmental impact analysis, and generating an environmental impact coefficient and a human skin color coefficient;

[0047] Step S400: establishing a deep learning model, analyzing and comparing the environmental impact coefficient and the human skin color impact coefficient to obtain an image processing mark;

[0048] Step S500: Analyze and generate false content identification, suspicious content identification and credible content identification based on the human skin color change mark and the image processing mark, and mark the matching identification interval on the video editing software.

[0049] It should be noted that preprocessing is performed through video editing software to perform operations such as denoising, image enhancement, and motion compensation.

[0050] Human skin color data includes facial skin color data X and body skin color data Y. It should be noted that human skin color data can be obtained through a skin color model, and the skin color model can be obtained using a skin color segmentation algorithm. The basic algorithms involved in the skin color segmentation algorithm include threshold segmentation, color space variation, and Gaussian mixture model segmentation methods.

[0051] Environmental impact data include ambient light intensity factor L, ambient reflection factor R, and light source angle factor F;

[0052] It should be noted that the environmental impact data is obtained through OpenCV. Specifically, the data is obtained through the following algorithm:

[0053] Light intensity factor L, histogram equalization image processing algorithm, by adjusting the image of a single identification interval to enhance the image contrast and brightness, and indirectly obtain light intensity factor data;

[0054] Environmental reflection factor R, normal estimation algorithm, estimates the normal direction of the object surface by analyzing the texture and shadow of the object surface in the image. The normal direction can be used to infer the angle factor of the light source;

[0055] Light source angle factor F, bilateral filtering, the bilateral filtering is an image smoothing algorithm that retains the edge information of the image while smoothing the color. The use of bilateral filtering on the image can suppress the influence of lighting changes, thereby estimating the environmental reflection factor.

[0056] The process of analyzing and generating human skin color change markers is as follows:

[0057] A human skin color comparison model is established. The human skin color comparison model can use a deep learning model, such as a convolutional neural network, to train the model to recognize and mark the human skin color in the image, and use the Fitzpatrick skin color classification method to assess the skin color level. Different skin colors are defined by numbers through assignment. The whiter the skin, the larger the corresponding skin level value.

[0058] The video to be identified is divided into n identification intervals, where n is an integer greater than 1. A single identification interval consists of two frames of images. Images of frames k and k+1 in the n identification intervals are obtained respectively. The facial skin color data comparison threshold reference values ​​X1 and X2 are set, where X1 < X2, and both X1 and X2 are greater than 0. The facial skin color data X1 at frame k is obtained. k and the facial skin color data X at frame k+1 k+1 , generate the facial skin color data difference Xc between frame k and frame k+1 through data processing, where X=X k+1 -X k (X k 、X k+1 is greater than 0, Xc is greater than or equal to 0), substituting the facial skin color data difference Xc into the facial skin color comparison threshold for comparative analysis to generate facial skin color level information, the facial skin color level information including primary facial skin color information, secondary facial skin color information and tertiary facial skin color information;

[0059] The logic for generating facial skin color level information is as follows: when the facial skin color data difference value Xc is less than the threshold reference value X1, the identification interval in which the facial skin color data difference value Xc is located is marked as first-level facial skin color information; when the facial skin color data difference value Xc is greater than the threshold reference value X1 and the facial skin color data difference value Xc is less than the threshold reference value X2, the identification interval in which the facial skin color data difference value Xc is located is marked as second-level facial skin color information; when the facial skin color data Xc is greater than the threshold reference value X2, the identification interval in which the facial skin color data difference value Xc is located is marked as third-level facial skin color information;

[0060] It should be noted that the identification interval corresponding to the facial skin color data difference Xc is the image of the k frame and the k+1 frame. The first-level facial skin color information is darker than the second-level facial skin color information, and the second-level facial skin color information is darker than the third-level facial skin color information. The whitening degree from the first-level facial skin color information to the third-level facial skin color information is from low to high.

[0061] Set the body skin color data comparison threshold reference values ​​Y1 and Y2, where Y1 < Y2, Y1 and Y2 are both greater than 0, and obtain the body skin color data Y at frame k. k and the body skin color data Y at frame k+1 k+1, generate the difference Yc of the body skin color data between frame k and frame k+1 through data processing, where Y=Y k+1 -Y k (Y k 、Y k+1 is greater than 0, Y is greater than or equal to 0), substituting the body skin color data difference Yc into the body skin color comparison threshold for comparative analysis to generate body skin color grade information, which includes primary body skin color information, secondary body skin color information, and tertiary body skin color information;

[0062] The logic for generating body skin color level information is as follows: when the body skin color data difference value Yc is less than the threshold reference value Y1, the identification interval of the body skin color data difference value Yc is marked as the first-level body skin color information; when the body skin color data is greater than the threshold reference value Y1 and the body skin color data difference value Yc is less than the threshold reference value Y2, the identification interval of the body skin color data difference value Yc is marked as the second-level body skin color information; when the body skin color data difference value Yc is greater than the threshold reference value Y2, the identification interval of the body skin color data difference value Yc is marked as the third-level body skin color information;

[0063] It should be noted that the identification interval corresponding to the body skin color data difference Yc is the image of frame k and frame k+1. The first-level body skin color information is darker than the second-level body skin color information, and the second-level body skin color information is darker than the third-level body skin color information. The whitening degree from the first-level body skin color information to the third-level body skin color information is from low to high.

[0064] The facial skin color level information and body skin color level information in a single identification area are integrated and analyzed to generate a human skin color change mark:

[0065] When primary facial skin color information and primary body skin color information exist simultaneously in a single identification area, a primary skin color change mark is generated for the identification area; when secondary facial skin color information and primary body skin color information, primary facial skin color information and secondary body skin color information, or secondary facial information and secondary body skin color information exist simultaneously in a single identification area, an intermediate skin color change mark is generated for the identification area; when tertiary facial skin color information and primary body skin color information, tertiary facial skin color information and secondary body skin color information, primary facial skin color information and tertiary body skin color information, secondary facial information and tertiary body skin color information, or tertiary facial information and tertiary body skin color information exist simultaneously in a single identification area, an advanced skin color change mark is generated for the identification area;

[0066] It should be noted that the advanced skin color change mark corresponds to the large difference in skin color changes between the two frames of images in the identification area, and there is a high probability of suspicion of video P-image forgery. Similarly, the probability of suspicion of P-image forgery corresponding to the advanced skin color change mark to the primary change mark is from high to low.

[0067] The steps to generate the environmental impact coefficient and human skin color coefficient are as follows:

[0068] A deep learning model is established to analyze and process environmental impact data. The analysis and processing logic is as follows:

[0069] The generation logic of the environmental impact coefficient is:

[0070] In the environmental data comparison model, a combined model of illumination error coefficients is designed, weights are assigned to each molecule, and the environmental impact coefficient δn is calculated based on the formula analysis. The specific formula is:

[0071] δn= ;

[0072] Where n is the nth identification interval, n is greater than 0, a, b, and c are the preset proportional coefficient values ​​of the light intensity factor L, the ambient reflection factor R, and the light source angle factor F, respectively, a>b>c>0, and a+b+c=1.425;

[0073] It should be noted that the algorithm involved in the illumination error coefficient combination model can adopt an illumination estimation algorithm. The global illumination estimation and spherical harmonic expansion in the illumination estimation algorithm can use physical models of illumination, such as light propagation models and reflection models, to infer environmental parameters by optimizing or solving a system of equations.

[0074] In addition, the larger the light intensity factor L, the greater the brightness of the environment, the more obvious the lighting on the human face, the larger the environmental reflection factor R, the higher the skin brightness of the portrait in the video, the larger the light source angle factor F, the more dispersed the lighting received by the person. When the light source angle is small, the light is relatively concentrated, resulting in clearer shadows and highlights. In this case, the details and features on the face may be more prominent, and the distribution of lighting is more obvious. When the light source angle is large, the light will shine on the face at a wider angle, resulting in a more uniform lighting distribution. In this case, the difference between shadows and highlights is smaller, the details on the face may be smoothed, and the distribution of lighting is more uniform. Analysis shows that the greater the light, the smaller the environmental impact, the larger the environmental impact coefficient δn, and the smaller the environmental impact, the smaller the environmental impact coefficient δn.

[0075] The generation logic of human skin color coefficient is:

[0076] The human skin color change mark includes a primary skin color change mark, an intermediate skin color change mark, and an advanced skin color change mark. A skin color distribution model is set in the environmental data comparison model. The facial skin color data X and the body skin color data Y are analyzed and processed in a formula to generate a human skin color coefficient βn. The human skin color coefficient βn is rooted by taking the ratio of the absolute value of the difference between the facial skin color data X and Y and Y and adding the skin color error ratio correction constant K. The specific formula of the human skin color coefficient βn is:

[0077] βn = (X, Y and K are all greater than 0, n is the nth identification interval, n>0);

[0078] It should be noted that the skin color distribution model uses a Gaussian mixture to model the skin color distribution. It is assumed that the skin color distribution is composed of multiple Gaussian distributions, each representing a different skin color component. By clustering the pixels in the image and estimating the parameters, a Gaussian mixture model of the skin color can be obtained. Subsequently, the characteristic function and latent variables of the skin color in the Gaussian mixture model are defined through the conditional random field probability model. The conditional probability distribution between the skin color data can be established, thereby realizing the analysis and inference of the skin color data;

[0079] In addition, the human skin color coefficient βn reflects the degree of influence of the human skin color. Under the same ambient lighting conditions, people with lighter skin colors are more likely to show changes in brightness under lighting conditions, while darker skin colors have relatively stable brightness under lighting conditions. The greater the difference between the facial skin color data X and the body skin color value Y, that is, the greater the difference between the facial skin color and the body skin color, the larger the human skin color coefficient βn will be. The smaller the difference between the facial skin color data X and the body skin color value Y, that is, the smaller the difference between the facial skin color and the body skin color, the smaller the human skin color coefficient βn will be.

[0080] The generation logic of image processing markup is:

[0081] The image processing mark includes a primary image processing mark, a secondary image processing mark, and a tertiary image processing mark. The environmental impact coefficient and the human skin color impact coefficient are analyzed and compared by formulating, and the influence of the environment in the image is removed to generate the original image coefficient γn. The specific formula is:

[0082] γn=( ) ( > > 0, q is the influence elimination error constant, q > 0, n is the nth identification interval, n > 0);

[0083] Set the pre-image coefficient comparison thresholds W1 and W2, where W1>W2>0, and substitute the pre-image coefficient γn into the pre-image coefficient comparison thresholds for analysis:

[0084] When the pre-image coefficient γn of a single identification area is less than the pre-image coefficient comparison threshold W2, a first-level image processing mark is generated for the identification area. When the pre-image coefficient γn of a single identification area is greater than the pre-image coefficient comparison threshold W2 and less than the pre-image coefficient comparison threshold W1, a second-level image processing mark is generated for the identification area. When the pre-image coefficient γn of a single identification area is greater than the pre-image coefficient comparison threshold W1, a third-level image processing mark is generated for the identification area.

[0085] It should be noted that the degree of influence of human skin color corresponding to the secondary image processing mark is higher than that of the primary image processing mark, and the degree of influence of human skin color corresponding to the tertiary image processing mark is higher than that of the secondary image processing mark.

[0086] Based on the human skin color change mark and image processing mark, false content identification, suspicious content identification and credible content identification are generated. The generation logic is as follows:

[0087] When a single identification area contains both a primary image processing mark and a mid-level skin color change mark, a primary image processing mark and a high-level skin color change mark, or a tertiary image processing mark and a primary skin color change mark, a false content mark is generated for the identification area;

[0088] When a secondary image processing mark and a primary skin color change mark, a secondary image processing mark and a high-level skin color change mark, a tertiary image processing mark and a mid-level skin color change mark, or a tertiary environmental impact mark and a high-level skin color change mark coexist within a single identification interval, a suspicious content mark is generated for the identification area;

[0089] When a single identification interval contains a primary image processing mark and a primary skin color change mark, a secondary image processing mark and a mid-level skin color change mark, or a tertiary image processing mark and a high-level skin color change mark, a trusted content identifier is generated for the identification area.

[0090] Images that match the identification interval are marked as false content, suspicious content, or credible content on the video editing software.

[0091] Through the above-mentioned implementation method, the identification video is analyzed frame by frame, and the skin color difference ratio between the human facial skin color and the body skin color is used to achieve more accurate video authenticity identification in the case of portrait photo retouching and P-picture. Through secondary processing of the first identification result, the environmental influencing factors and the obtained first identification result are analyzed and processed for further judgment. In the process of authenticity identification of video content with portraits, the accuracy of the identification results and the diversity of identification methods are improved.

[0092] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0093] It should be understood that the term "and / or" as used herein simply describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. A and B can be singular or plural. Furthermore, the character " / " as used herein generally indicates an "or" relationship between the associated objects, but it may also indicate an "and / or" relationship. For specific understanding, please refer to the context.

[0094] In this application, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, "at least one of a, b, or c" can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural.

[0095] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0096] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0097] The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, and may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment as needed.

[0098] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0099] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. A method for identifying the authenticity of video content, characterized in that: The method comprises the following steps: Step S100: Preprocessing the video to be identified and obtaining video content identification data frame by frame. The video content identification data includes human skin color data and environmental impact data. Step S200: establishing a human skin color comparison model, substituting the acquired human skin color data into the human skin color comparison model to perform human skin color analysis to obtain a human skin color change mark; The process of analyzing and generating human skin color change markers is as follows: The video to be identified is divided into n identification intervals, each of which consists of two frames of images. Images of frames k and k+1 within the n identification intervals are obtained respectively. The facial skin color data difference Xc between frames k and k+1 is generated through data processing. The facial skin color data difference Xc is substituted into the facial skin color comparison threshold for comparative analysis to generate facial skin color level information; Generate the body skin color data difference Yc between frame k and frame k+1 through data processing, substitute the body skin color data difference Yc into the body skin color comparison threshold for comparative analysis, and generate body skin color level information; Step S300: establishing an environmental data comparison model, substituting the acquired human skin color data and environmental impact data into the environmental data comparison model to perform environmental impact analysis, and generating an environmental impact coefficient and a human skin color coefficient; The generation logic of the environmental impact coefficient is: In the environmental data comparison model, a combined model of illumination error coefficients is designed, weights are assigned to each molecule, and the environmental impact coefficient δn is calculated based on the formula analysis. The specific formula is: δn= ; Where n is the nth identification interval, n is greater than 0, a, b, and c are the preset proportional coefficient values ​​of the light intensity factor L, the ambient reflection factor R, and the light source angle factor F, respectively, a>b>c>0, and a+b+c=1.425; The generation logic of human skin color coefficient is: The human skin color change mark includes a primary skin color change mark, an intermediate skin color change mark, and an advanced skin color change mark. A skin color distribution model is set in the environmental data comparison model. The facial skin color data X and the body skin color data Y are analyzed and processed in a formula to generate a human skin color coefficient βn. The human skin color coefficient βn is rooted by taking the ratio of the absolute value of the difference between the facial skin color data X and Y and Y and adding the skin color error ratio correction constant K. The specific formula of the human skin color coefficient βn is: βn = ; X, Y and K are all greater than 0, n is the nth identification interval, n>0; Step S400: Analyze and compare the environmental impact coefficient and the human skin color impact coefficient to obtain an image processing mark; The generation logic of image processing markup is: By formulating and analyzing the environmental impact coefficient and the human skin color impact coefficient, the influence of the environment in the image is removed and the original image coefficient γn is generated. The specific formula is: γn=( ) ; > >0, q is the error constant affecting elimination, q>0, n is the nth identification interval, n>0; Step S500: Analyze and generate false content identification, suspicious content identification and credible content identification based on the human skin color change mark and the image processing mark, and mark the matching identification interval on the video editing software.

2. A method for identifying the authenticity of video content according to claim 1, characterized in that: The human skin color data includes facial skin color data X and body skin color data Y, and the environmental impact data includes an environmental light intensity factor L, an environmental reflection factor R, and a light source angle factor F.

3. A method for identifying the authenticity of video content according to claim 2, characterized in that: The facial skin color level information generation logic is: When the facial skin color data difference value Xc is less than the threshold reference value X1, the identification interval where the facial skin color data difference value Xc is located is marked as the first-level facial skin color information; when the facial skin color data difference value Xc is greater than the threshold reference value X1 and the facial skin color data difference value Xc is less than the threshold reference value X2, the identification interval where the facial skin color data difference value Xc is located is marked as the second-level facial skin color information; when the facial skin color data Xc is greater than the threshold reference value X2, the identification interval where the facial skin color data difference value Xc is located is marked as the third-level facial skin color information; The logic for generating the body skin color level information is as follows: When the body skin color data difference Yc is less than the threshold reference value Y1, the identification interval where the body skin color data difference Yc is located is marked as the first-level body skin color information; when the body skin color data is greater than the threshold reference value Y1 and the body skin color data difference Yc is less than the threshold reference value Y2, the identification interval where the body skin color data difference Yc is located is marked as the second-level body skin color information; when the body skin color data difference Yc is greater than the threshold reference value Y2, the identification interval where the body skin color data difference Yc is located is marked as the third-level body skin color information.

4. A method for identifying the authenticity of video content according to claim 3, characterized in that: Based on the human skin color change mark and image processing mark, false content identification, suspicious content identification and credible content identification are generated. The generation logic is as follows: When a single identification area contains both a primary image processing mark and a mid-level skin color change mark, a primary image processing mark and a high-level skin color change mark, or a tertiary image processing mark and a primary skin color change mark, a false content mark is generated for the identification area; When a secondary image processing mark and a primary skin color change mark, a secondary image processing mark and a high-level skin color change mark, a tertiary image processing mark and a mid-level skin color change mark, or a tertiary environmental impact mark and a high-level skin color change mark coexist within a single identification interval, a suspicious content mark is generated for the identification area; When a single identification interval contains a primary image processing mark and a primary skin color change mark, a secondary image processing mark and a mid-level skin color change mark, or a tertiary image processing mark and a high-level skin color change mark, a trusted content identifier is generated for the identification area.

Images