Batch picture authenticity analysis system and method
By designing a batch image authenticity analysis system, using the verification value and metadata information of image files for classification and screening, the problem of inefficient image authenticity identification in the existing technology is solved, and automated processing and targeted basis output are realized. Improved identification efficiency and coverage.
Patent Information
- Application Number
- CN202411261396.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-10
- Publication Date
- 2024-12-03
AI Technical Summary
The prior art lacks a complete judgment system and adaptive selection when processing the authenticity of a large number of pictures, resulting in low processing efficiency and inability to obtain targeted conclusions.
A batch image authenticity analysis system was designed, including input module, classification module, basis extraction module, initial screening module and comparison module. By calculating the verification value of image files, extracting time and GPS information in metadata, preliminary work is carried out. Filter and compare, output targeted basis, and use SHA256 verification algorithm and EXIF information to classify and judge images.
It realizes automated batch processing, improves the efficiency of picture authenticity identification, reduces manual operations, expands the channels for feature acquisition, provides targeted and objective basis, and helps inspectors to allocate their work efficiently.
Smart Images

Figure CN119068294A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of judicial identification, and in particular to a batch picture authenticity analysis system and method based on EXIF information. Background Art
[0002] With the progress of society and the rapid development of modern image technology, digital photos have gradually penetrated into all aspects of people's daily work and life, and are playing an increasingly important role. The popularity of portable devices such as mobile phones, tablets, and watches enables ordinary people to take photos at any time. With the widespread use of photos as evidence in various cases, a large number of problems have also arisen. In criminal cases and civil disputes, a series of problems related to whether the pictures have been edited are often encountered, such as extracting the formation and modification time of the picture record, determining the shooting location of the picture, determining whether the picture has been tampered with, determining the relationship between pictures, etc. These problems have put forward new requirements for image identification technology.
[0003] In the prior art, there have been some studies on various features of images, but these studies have not formed a complete judgment system, and lack adaptive selection and batch processing rules for actual cases, resulting in low efficiency in processing the authenticity of images in actual cases and inability to obtain targeted conclusions. Summary of the invention
[0004] In order to solve the above technical problems, the present invention provides a batch image authenticity analysis system and method, which solves the problem of outputting targeted evidence and batch processing when authenticity identification of a large number of images in actual cases.
[0005] The technical solution adopted by the present invention to solve the technical problem is: A batch image authenticity analysis system, specifically comprising: The input module is used to read all the image files and their basic information of the current inspection, and output the basic information of the image files to the output module.
[0006] A classification module, connected to the input module, is used to calculate the uniqueness check value of all input image files and classify them. The extraction result of the classification module is output to the basis extraction module and also to the output module; The classification module specifically includes: a check value calculation unit, which is used to calculate the check values of all image files checked at the time; a classification unit, which is used to classify the consistency of the image according to the check value calculation results, classify image files with the same check value into the same category, and output the results to the output module at the same time.
[0007] According to the extraction module, the classification module is connected to extract key information related to authenticity judgment in the metadata of each image to be inspected.
[0008] The initial screening module is connected to the basis extraction module and is used to make a preliminary judgment on the result of the basis extraction module of each image, and obtain the preliminary screening result of the image with obvious authenticity defects in the image metadata, and output the result to the output module.
[0009] The comparison module is connected to the basis extraction module and is used to compare the results of the basis extraction modules between different images, and obtain the preliminary screening results of images with obvious authenticity defects between the images, and output the results to the output module.
[0010] The output module is connected to the above-mentioned input module, classification module, basis extraction module, preliminary screening module and comparison module, and is used to store the processing results of the above-mentioned modules and sort and output them.
[0011] Furthermore, the basic information of the image file includes the file name and the file path.
[0012] Furthermore, the key information related to authenticity judgment includes time information such as creation time and modification time stored in the image metadata, and GPS information.
[0013] Furthermore, the image metadata has obvious authenticity flaws, including the image lacking necessary time information, image editing software watermarks stored in the image metadata, image modification time being later than creation time, and multiple time information for the same image having excessive differences.
[0014] Furthermore, there are obvious authenticity defects between the images, including that the GPS locations of multiple images with different path sources are consistent, and the modification times of the images are consistent or very close.
[0015] A batch image authenticity analysis method specifically includes the following steps: Step S1. Batch extract all images to be inspected: the batch image authenticity analysis system batch imports all images to be inspected, and simultaneously obtains basic information such as file names and paths of all images to be inspected; Step S2. Check value calculation: The system calculates and classifies all input image files, and classifies image files with the same check value into the same category; Step S3. Authenticity extraction: The system extracts key information related to authenticity judgment from the metadata of each image to be inspected, including time, GPS information, etc., and outputs the results to the output module at the same time; Step S4. Preliminary screening of extracted data: the system makes a preliminary judgment on each image based on the results of the extraction module, processes and obtains preliminary screening results of images with obvious authenticity defects in the image metadata, and outputs the results to the output module; Step S5. Extraction data comparison: The system compares the results of the extraction module between different images, processes and obtains preliminary screening results of images with obvious authenticity defects, and outputs the results to the output module; Step S6. Output to table: The system is used to store the processing results of the above modules and output them in order.
[0016] Furthermore, in step S2, the classification specifically includes: first performing a check value calculation to calculate the check values of all image files that are inspected at the time, and then performing classification, classifying the consistency of the image according to the check value calculation results, classifying image files with the same check value into the same category, and outputting the results to the output module at the same time.
[0017] Furthermore, in the step S2, the check value calculation, it is recommended to use the SHA256 check algorithm.
[0018] Furthermore, in step S3, the key information related to authenticity judgment includes time information and GPS information stored in the image metadata.
[0019] Furthermore, in step S3, the time information includes access time, creation time, file modification time, generation time and modification time stored in metadata related to file system attributes.
[0020] Furthermore, in step S3, the GPS information includes longitude, latitude, altitude, and shooting direction.
[0021] Furthermore, in step S4, the image metadata has obvious authenticity flaws, including the image lacking necessary time information, image editing software watermarks stored in the image metadata, image modification time being later than creation time, and multiple time information of the same image having excessive differences.
[0022] Furthermore, in step S4, the image editing software watermark refers to the representation information written by the later image editing software in the image EXIF information, including the name of the editing software and also includes a descriptive statement for recording the editing event.
[0023] Furthermore, in step S5, the images have obvious authenticity defects, including that the GPS locations of multiple images with different path sources are consistent, the modification times of the images are consistent or very close, and multiple images taken by the same model of equipment have different data structures and sizes.
[0024] The beneficial effects of the invention are: The above technical solution can solve the problem of outputting targeted evidence and batch processing when authenticating a large number of pictures, as follows: 1. Automated batch processing: The system of the present invention can automatically process a large number of images, significantly improving the efficiency of image authenticity identification. At the same time, it also reduces the need for manual operation and the possibility of misjudgment caused by human factors.
[0025] 2. Comparison between images: The present invention not only verifies the authenticity features within the image, but also takes into account the comparison between images submitted for inspection in the same batch, thus expanding the channels for feature acquisition and improving the coverage of authenticity identification inspection.
[0026] 3. Targeted basis output: The output of the present invention covers the basic features in authenticity identification, and at the same time performs logical judgment on the features with higher value, and outputs them separately after preliminary screening, so as to provide objective basis for inspection in a focused manner, and at the same time, help inspectors to allocate working time in a targeted manner. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] The present invention will be further described below in conjunction with the accompanying drawings.
[0028] Figure 1 The figure is a diagram of the module composition and relationship of the present invention.
[0029] Figure 2 The figure is a flow chart of batch picture authenticity analysis in a preferred embodiment of the present invention.
[0030] Figure 3 The present invention is a flowchart for refining the authenticity analysis of batch images in a preferred embodiment of the present invention.
[0031] Figure 4 It is a table structure diagram of the present invention. DETAILED DESCRIPTION
[0032] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. The present invention is not limited to this embodiment, and other embodiments may also fall within the scope of the present invention as long as they conform to the gist of the present invention.
[0033] like Figure 1 As shown, the present invention provides a batch image authenticity analysis system, including an input module, a classification module, a basis extraction module, a preliminary screening module, a comparison module, and an output module. The classification module specifically includes: a check value calculation unit and a classification unit. After the image file data is input into the system from the input module, it is sequentially transmitted through the classification module and the basis extraction module, and then transmitted to the preliminary screening module and the comparison module. At the same time, the processing results of each module are output to the output module. The modules are interconnected to form a complete analysis process.
[0034] In a preferred embodiment of the present invention, based on the above problems existing in the prior art, a batch image authenticity analysis system is provided. Figure 2shown.
[0035] The detailed process is as follows Figure 3 As shown, first, the file names of all the images to be inspected are exported in batches, output as A0, and the paths are output as A1, and stored separately to avoid conflicts between files with the same name in different paths.
[0036] Then calculate the verification value of each image to be checked. The SHA256 verification algorithm can be used to calculate the SHA256 hash value output as A2.
[0037] All the checksum values are compared and all the images are divided into two categories: one is images with the same hash value, and the other is images with unique hash values. For images with unique hash values, their A3 key value is assigned 0. For images with non-unique hash values, they are grouped by hash value, with hash value A2 as the group number. At the same time, the number of images in each group A3 is calculated.
[0038] Extract key authenticity-related information from the metadata of each image to be inspected, such as creation, modification, generation and other time information, which are stored as T1 to Tn respectively, GPS domain time information is stored as TG, GPS domain positioning information is stored as A5, generation software information is stored as A6, shooting equipment and model is stored as A7, editing software information is stored as A8, etc.
[0039] The extracted data is preliminarily screened for authenticity features, including the maximum time difference between T1 and Tn A9, the time difference between the time information in the GPS domain and T1 to Tn A10, the difference between the shooting equipment and model information and the generated software information, and the matching degree between the editing software information and the suspicious editing software library A12.
[0040] Then, the information differences between the images to be inspected are compared, such as the generation time difference between the images A13, the GPS positioning difference A14, and the image content difference using the similarity of the picture elements for comparison and storage as A15.
[0041] Finally, the above results are output in descending order of the number of groups, time difference, suspicious database matching degree, and hash value, ensuring that the pictures with the largest anomalies and high suspicion appear at the front of the output results. The table structure used to store the final output results is as follows: Figure 4 As shown, specifically: A0 The name of the file to be checked includes two parts: the name and the extension. Due to the rules of the computer operating system, files in the same folder cannot have the same name.
[0042] A1: The storage location of the file to be checked, excluding the file name.
[0043] A2 integrity check value transforms input data of any length into an output value of fixed length through a specific hash algorithm, which is used to identify the uniqueness or integrity of the data. Commonly used hash algorithms include MD5, SHA256, etc.
[0044] A3 When the hash value of a file is not unique among all the files to be checked, a group number is assigned starting from 1 according to the number of files in the group A4 and the hash value A2. When the hash value is unique, the group number is assigned 0. Whether the group number is 0 is also a key value for distinguishing whether the file is repeated. When the file should not be repeated, the file with the group number 0 will enter the subsequent judgment. The file with the group number not 0 is directly judged to be suspicious.
[0045] A4 is the number of times a certain hash value appears, used to identify the number of times a duplicate file appears.
[0046] A5 is extracted from EXIF information, Tag=0x8825 in IFD0 (main image file directory). If the shooting device has GPS function, A5 records longitude and latitude data. If the shooting device does not have GPS function, A5 is assigned "(empty)". In some image requirements used to prove the shooting location, when the A5 value is empty, the GPS positioning data is manually reviewed.
[0047] A6 is extracted from EXIF information, Tag=0x0131 in IFD0, and null value is assigned "(null)" A7 is extracted from EXIF information, concatenating the contents of Tag=0x010F and 0x0110 in IFD0. If both are null values, assign "(null)" A8 is extracted from EXIF information, Tag=0x000b or 0xa43b in IFD0, and null value is assigned "(null)" A9 The maximum difference between multiple attribute times. Select the maximum and minimum values under items containing time information such as DateTimeDigitized, DateTime, and DateTimeOriginal and calculate their difference.
[0048] A10The difference between GPS time and the maximum value of attribute time. Select the maximum value under the items containing time, such as creation time, ModifyDate modification date, original time, etc., and calculate the difference with GPS time.
[0049] A11 The difference between the creation software and the editing software recorded in the EXIF information. When the contents of the two are inconsistent, this item is expressed as "creation software" VS "editing software"; when the contents of the two are consistent, this item is assigned "(consistent)".
[0050] The editing software recorded in the A12EXIF information hits the image processing tool watermark information in the suspicious library.
[0051] The suspicious library is composed of pre-acquired watermark information of image editing software. The suspicious library includes common known intuitive watermarks such as "Photoshop", "CorelDRAW", and "ACDSee", as well as watermark information such as "edit" and "history" that may be written by editing software.
[0052] A13 selects the maximum and minimum values of DateTimeOriginal in a set of files to be inspected and calculates the difference. A14 selects the maximum and minimum values of GPS in a set of files to be inspected and calculates the difference.
[0053] A15 compares the differences between feature vectors of a set of files to be inspected and evaluates the similarity of the images.
[0054] TG is extracted from EXIF information, Tag=0x8825 in IFD0, and empty value is assigned "(empty)".
[0055] T1-Tn assigns items containing time information such as DateTimeDigitized, DateTime, and DateTimeOriginal to T1-Tn one by one.
[0056] Those skilled in the art may make various modifications and changes to the present invention. Therefore, the present invention covers various modifications and changes that fall within the scope of the appended claims and their equivalents.
Claims
1. A batch image authenticity analysis system, specifically include: The input module is used to read all the image files and their basic information of the current inspection, and output the basic information of the image files to the output module; A classification module, connected to the input module, is used to calculate the uniqueness check value of all input image files and classify them. The extraction result of the classification module is output to the basis extraction module and also to the output module; The classification module specifically includes: a check value calculation unit, which is used to calculate the check values of all image files that are checked at the time, and a classification unit, which is used to classify the consistency of the image according to the check value calculation result, classify the image files with the same check value into the same category, and output the results to the output module at the same time; According to the extraction module, the classification module is connected to extract key information related to authenticity judgment in the metadata of each image to be inspected; A preliminary screening module, connected to the basis extraction module, is used to make a preliminary judgment on the result of the basis extraction module of each image, and obtain the preliminary screening result of the image with obvious authenticity defects in the image metadata, and output the result to the output module; A comparison module, connected to the basis extraction module, for comparing the results of the basis extraction modules between different images, processing to obtain a preliminary screening result of images with obvious authenticity defects between the images, and outputting the result to the output module; The output module is connected to the above-mentioned input module, classification module, basis extraction module, preliminary screening module and comparison module, and is used to store the processing results of the above-mentioned modules and sort and output them.
2. A batch image authenticity analysis system according to claim 1, It is characterized in that The basic information of the image file includes the file name and the file path.
3. A batch image authenticity analysis system according to claim 1, It is characterized in that The key information related to authenticity judgment includes time information such as creation time and modification time stored in the image metadata, and GPS information.
4. A batch image authenticity analysis system according to claim 1, It is characterized in that The image metadata has obvious authenticity flaws, including the image lacking necessary time information, image editing software watermarks stored in the image metadata, image modification time being later than creation time, and multiple time information items of the same image having excessive differences.
5. A batch image authenticity analysis system according to claim 1, It is characterized in that The images have obvious authenticity defects, including that the GPS locations of multiple images with different path sources are consistent, and the modification times of the images are consistent or very close.
6. A batch image authenticity analysis method, applied to the batch image authenticity analysis system as claimed in any one of claims 1 to 5, specifically The following steps are involved: Step S1. Batch extract all images to be inspected: the batch image authenticity analysis system batch imports all images to be inspected, and simultaneously obtains basic information such as file names and paths of all images to be inspected; Step S2. Check value calculation: The system calculates and classifies all input image files, and classifies image files with the same check value into the same category; Step S3. Authenticity extraction: The system extracts key information related to authenticity judgment from the metadata of each image to be inspected, including time, GPS information, etc., and outputs the results to the output module at the same time; Step S4. Preliminary screening of extracted data: the system makes a preliminary judgment on each image based on the results of the extraction module, processes and obtains preliminary screening results of images with obvious authenticity defects in the image metadata, and outputs the results to the output module; Step S5. Extraction data comparison: The system compares the results of the extraction module between different images, processes and obtains preliminary screening results of images with obvious authenticity defects, and outputs the results to the output module; Step S6. Output to table: The system is used to store the processing results of the above modules and output them in order.
7. A batch image authenticity analysis method according to claim 6, It is characterized in that In step S2, the classification specifically includes: first performing a check value calculation to calculate the check values of all image files that are checked at the time, and then performing classification, classifying the consistency of the image according to the check value calculation results, classifying image files with the same check value into the same category, and outputting the results to the output module at the same time; for the check value calculation, it is recommended to use the SHA256 check algorithm.
8. A batch image authenticity analysis method according to claim 6, It is characterized in that In step S3, the key information related to authenticity judgment includes time information and GPS information stored in image metadata; the time information includes access time, creation time, file modification time related to file system attributes, generation time and modification time stored in metadata; the GPS information includes longitude, latitude, altitude, and shooting direction.
9. A batch image authenticity analysis method according to claim 6, It is characterized in that In step S4, the image metadata has obvious authenticity flaws, including the image lacking necessary time information, image editing software watermark stored in the image metadata, image modification time being later than creation time, and multiple time information of the same image differing too much; the image editing software watermark refers to the representation information written by the later image editing software in the image EXIF information, including the name of the editing software, and also includes a descriptive statement for recording the editing event.
10. A batch image authenticity analysis method according to claim 6, It is characterized in that In the step S5, the images have obvious authenticity defects, including that the GPS locations of multiple images with different path sources are consistent, the modification time of the images is consistent or very close, and multiple images taken by the same model of equipment have different data structures and sizes.
Citation Information
Patent Citations
Archive auditing method and device, computer equipment and storage medium
CN112613340A
Image tampering detection method and processor
CN115527101A
Method and device for detecting and checking duplicate of forged image
CN116205850A
Tampering detection apparatus and tampering detection method
JP2010154456A
