An artificial intelligence-based computer vision image processing method

Abstract

The application discloses a computer vision image processing method based on artificial intelligence, receives a portrait image to be processed and an image processing task number, obtains business purpose identification and regional processing authorization information, divides image regions by using a computer vision model group, generates a regional processing authority table according to the business purpose identification, the regional processing authorization information, a purpose processing rule table and a region type, performs differentiated processing according to regions, forms a processing result set bound with a region source marker and a data retention control marker, reorganizes a purpose isolation processing result according to a purpose region mapping relationship, performs output verification according to the regional processing authority table, the region source marker and the data retention control marker, performs irreversible processing on an image region still corresponding to restricted original region data and updates a result, and generates a regional processing record after verification. The application can improve the controllability and traceability of portrait image processing under multiple business purposes.

Description

Technical Field

[0001] This invention relates to the field of computer vision image processing and data compliance management technology, and in particular to a computer vision image processing method based on artificial intelligence. Background Technology

[0002] Facial images are commonly used in business processes such as document submission, headshot display, background archiving, manual review, and document filing. The processing objectives for the same facial image to be processed differ in different business processes: for example, in document submission scenarios, it is usually necessary to preserve the clarity of the facial area; in headshot display scenarios, cropping or generating a display version may be necessary; and in background archiving or document filing scenarios, it may be necessary to restrict the storage of raw data in the background area, accessory areas, or other unnecessary areas.

[0003] Existing portrait image processing methods mostly adopt a uniform approach to processing the entire image. For operations such as portrait enhancement, image cropping, background blurring, or pixelation, the entire image or a fixed selected area is usually treated as the processing target. It is difficult to determine the processing method separately for the different attributes of the face area, hairstyle area, shoulder and neck area, background area, and accessory area. When different areas in the same image have different processing requirements, uniform processing of the entire image can easily lead to the enhancement, preservation, or output of non-target areas, and may also result in areas that need to be protected being displayed, stored, or written to the image library in the form of the original area data.

[0004] While some existing methods can segment human images using computer vision models, the segmentation results are typically only used for recognition, detection, enhancement, or background replacement, failing to form a stable data processing chain with business purpose identification, region processing authorization information, output control methods, and data retention control methods. When the same human image corresponds to multiple business purposes, existing methods struggle to generate application-isolated processing results based on different business purposes. They also struggle to record the spatial extent, processing basis, actual image processing method, region data source, and deletion status of original region cache data for each image region, making it difficult to subsequently verify and trace the image processing results under different business purposes.

[0005] Therefore, the technical problem that needs to be solved is how to use artificial intelligence models to obtain image regions that can be processed independently during computer vision image processing, and how to associate region coordinates, region masks, business purpose identifiers and region processing authorization information to determine image processing methods, output control methods and data retention control methods for different image regions under different business purposes, generate corresponding purpose isolation processing results, and determine whether there are still image regions in the purpose isolation processing results that are prohibited from being retained, displayed or stored in their original form and that correspond to the original region data. Summary of the Invention

[0006] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.

[0007] Therefore, to solve the above-mentioned technical problems, the present invention provides the following technical solution: a computer vision image processing method based on artificial intelligence, comprising the following steps performed sequentially: Receive the portrait image to be processed and the image processing task number; The business purpose identifier and regional processing authorization information are obtained based on the image processing task number, and the portrait image to be processed, the image processing task number, the business purpose identifier, and the regional processing authorization information are associated to form an image processing input data packet; The image of the person to be processed is input into a pre-trained computer vision model group to perform image region segmentation and generate image region segmentation results. The image region segmentation results include the region identifier, region type, region coordinates and region mask of each image region. Among them, for image regions that are adjacent or partially overlap, the ownership of overlapping pixels is determined according to a preset region ownership rule. Based on the business purpose identifier, area processing authorization information, purpose processing rule table, and area type in the image area division result, an area processing permission table corresponding to each image area is generated. The area processing permission table includes area identifier, area type, area coordinates, area mask, allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priority. Based on the region processing permission table and image region division results, region-differentiated processing is performed on each image region based on region coordinates and the scope of action limited by region masks, generating a set of region processing results. Each region processing result in the set is bound to a corresponding region identifier, actual image processing method, region source marker, and data retention control marker. The region source marker is used to identify the data source type of the corresponding region processing result, and the data retention control marker is used to identify the data retention control method applicable to the corresponding image region. Based on the business purpose identifier and the preset purpose area mapping relationship, the corresponding area processing result is read from the area processing result set, and the purpose isolation processing result is generated by recombining the area coordinates and area mask. Based on the region processing permission table, region source marker, and data retention control marker, the purpose isolation processing result is output verified to determine whether the image region that is prohibited from being retained, displayed, or stored in its original form still corresponds to the original region data. When the determination result is yes, at least one irreversible processing of blurring, masking, pixelation, or down-fidelity is re-executed on the corresponding image region, and the purpose isolation processing result is updated based on the re-processed region processing result. When the purpose isolation processing result does not contain any image regions that are prohibited from being retained, displayed, or stored in their original form but still correspond to the original region data, a region processing record associated with the purpose isolation processing result is generated.

[0008] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, the business purpose identifier is generated by the business system when creating an image processing task based on the business page entry, business process node or business submission type, and stored in correspondence with the image processing task number. The regional processing authorization information is generated by the business system based on the business purpose identifier, business scenario configuration and user authorization result, and is stored in correspondence with the image processing task number. The region processing authorization information includes at least the region type, authorized image processing methods, prohibited image processing methods, permitted output control methods, prohibited output control methods, permitted data retention control methods, and prohibited data retention control methods.

[0009] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, the method further includes the following steps before performing image region segmentation: reading image compliance parameters from a preset parameter table, the image compliance parameters including a preset image format, a preset pixel size range, a preset resolution range, and a preset file size range; performing image format verification, pixel size verification, resolution verification, and file size verification on the portrait image to be processed according to the image compliance parameters; performing image region segmentation when the image compliance verification result is passed; and generating an image compliance exception record when the image compliance verification result is failed.

[0010] As a preferred embodiment of the artificial intelligence-based computer vision image processing method of the present invention, the computer vision model group includes a pre-trained portrait semantic segmentation model, a facial key point detection model, a head and shoulder contour segmentation model, and an accessory detection model, and the image region division step includes: The human image to be processed is input into the human semantic segmentation model, which outputs a human subject mask and a non-human subject mask. The human subject region is determined based on the human subject mask, and the non-human subject region is determined based on the non-human subject mask. Within the main human figure area, the face region, hairstyle region, and shoulder and neck region are divided based on the facial key points output by the facial key point detection model and the head contour, hairstyle contour, and shoulder and neck contour output by the head and shoulder contour segmentation model. Define the background area within the non-human subject area; The model inputs the human image to be processed into the accessory detection model and outputs the accessory category, detection box, accessory confidence score, and accessory mask. When the accessory confidence score reaches the preset confidence threshold, the corresponding accessory is divided into one or more accessory regions. Generate corresponding region coordinates based on the region mask of each image region, and generate region identifiers and region types for each image region; For image regions that are adjacent to or partially overlap each other, the region assignment of overlapping pixels is determined according to the preset region priority. The preset region priority is set from high to low in the order of accessory region, face region, hairstyle region, shoulder and neck region and background region.

[0011] In a preferred embodiment of the artificial intelligence-based computer vision image processing method of the present invention, the step of generating the region processing permission table includes: Based on the business purpose identifier and region type, the corresponding purpose processing rules are read from the purpose processing rule table. The purpose processing rules include the default allowed image processing method, the default prohibited image processing method, the default output control method, the default data retention control method, and the default processing priority. The usage processing rules are matched with the regional processing authorization information; When either the usage processing rule or the area processing authorization information prohibits a certain image processing method, the image processing method is written into the prohibited image processing methods; When the application processing rule allows a certain image processing method and the region processing authorization information does not prohibit the image processing method, the image processing method is written into the allowed image processing methods. When any of the aforementioned usage processing rules and regional processing authorization information prohibits display in its original form, prohibits the entry of original regional data into the database, prohibits storage in its original form, or requires the deletion of original regional cache data, the corresponding control content will be written into the output control method or data retention control method. The region identifier, region type, region coordinates, region mask, allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priority are summarized to form a region processing permission table.

[0012] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, the image processing method includes at least one of the following: preservation, enhancement, cropping, display version generation, blurring, masking, pixelation, and fidelity reduction. The output control method includes at least one of allowing display, prohibiting display in its original form, allowing output, and prohibiting output. The data retention control method includes at least one of the following: allowing storage, prohibiting storage in its original form, prohibiting the entry of original area data into the database, and deleting original area cache data. When both image processing and data retention control methods exist in the same image area, the data retention control method is executed first to restrict the storage, entry into the database, or cache deletion of the original area data, and the actual image processing method is determined under the condition that the data retention control method is satisfied. When the allowed image processing methods for the same image region include enhancement as well as at least one of blurring, masking, pixelation, or defidelity, the actual image processing method shall be determined first from blurring, masking, pixelation, or defidelity. When there is an image region in the region processing permission table that has no processing permission configured, a default processing control method is configured for the image region according to the region type and business purpose identifier. The default processing control method includes at least one of the following: masking, pixelation, down-fidelity, prohibiting display in the original form, or prohibiting the original region data from being entered into the database.

[0013] In a preferred embodiment of the artificial intelligence-based computer vision image processing method of the present invention, the region differentiation processing includes: For image regions that are preserved in their original form and are not prohibited from being preserved, displayed or stored in their original form, the original region data of the image region is copied according to the corresponding region coordinates and region mask to generate the preserved region processing result. For image regions where the actual image processing method is enhancement, brightness equalization, noise suppression, edge preservation, or color correction are performed based on the corresponding region coordinates and region mask to generate the enhancement region processing result. For image regions generated by actual image processing methods such as cropping or display, perform boundary cropping, size adaptation, or display version generation processing based on the corresponding region coordinates and region mask to generate cropped region processing results or display version region processing results. For image regions that are actually blurred, masked, pixelated, or de-fidelity processed, irreversible processing is performed based on the corresponding region coordinates and region mask to generate irreversible region processing results.

[0014] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, the following is provided: for image regions where the data retention control method prohibits the entry of original region data into the database, the original region data corresponding to the image region is not written into the formal image database; when the business application requires the retention of the corresponding region results, only the irreversibly processed region data or the processing result index is written into the result database or record database. For image regions where the data retention control method is to delete the original region cache data, a cache mark to be deleted is generated after the purpose isolation processing result is generated, and the cache mark to be deleted is associated with the region identifier of the corresponding image region; when generating the region processing record, the cache mark to be deleted is written into the region processing record; after the corresponding original region cache data is deleted, the cache mark to be deleted in the region processing record is updated to the cache deletion completed mark.

[0015] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, each region processing result in the region processing result set is bound to the corresponding region identifier, region type, actual image processing method, processing result index, region source marker, and data retention control marker. The region source marker is used to distinguish whether the corresponding region processing result comes from the original region data, enhanced region data, cropped region data, display version region data, or irreversibly processed region data. When generating the purpose isolation processing result, the target business purpose version is determined according to the business purpose identifier, the region processing result corresponding to the target business purpose version is read according to the preset purpose region mapping relationship, and the image is reconstructed according to the region coordinates and region mask. When the region masks corresponding to multiple image regions overlap, the coverage relationship of overlapping pixels is determined according to the preset region priority.

[0016] As a preferred embodiment of the computer vision image processing method based on artificial intelligence described in this invention, after generating the purpose isolation processing result, the region source mark and data retention control mark of each region in the purpose isolation processing result are read, and it is determined according to the region processing permission table whether the image region that is prohibited from being retained, displayed or stored in its original form still corresponds to the original region data. When the initial judgment result is yes, at least one irreversible processing of blurring, masking, pixelation or downfidelity reduction is re-executed on the corresponding image region, and the purpose isolation processing result is updated based on the re-processed region processing result. The region source tag and data retention control tag of the corresponding image region in the updated purpose isolation processing result are read, and the region processing permission table is used to determine again whether the corresponding image region still corresponds to the original region data. If the result is yes again, generate an exception record for isolation processing. If the initial judgment result is negative, or if the second judgment result is negative, a region processing record is generated. The region processing record includes region identifier, region type, region coordinates, business purpose identifier, allowed image processing method, prohibited image processing method, output control method, data retention control method, processing priority, actual image processing method, processing result index, region source marker, data retention control marker, and cache deletion marker.

[0017] The beneficial effects of this invention are: 1. The present invention inputs the portrait image to be processed into a pre-trained computer vision model group, and divides the face region, hairstyle region, shoulder and neck region, background region and accessory region through portrait semantic segmentation, facial key point detection, head and shoulder contour segmentation and accessory detection, and obtains the region coordinates and region mask of each image region; thereby, subsequent image processing can be limited to the corresponding pixel range, reducing the possibility of misprocessing non-target regions.

[0018] 2. This invention generates a region processing permission table based on the business purpose identifier, region processing authorization information, purpose processing rule table, and image region division results, and binds the region processing permission table to region coordinates and region mask. Compared with the unified processing method for the whole image, this invention can determine the image processing method, output control method, and data retention control method for different image regions under different business purposes.

[0019] 3. This invention separates the image processing method from the data retention control method. The image processing method controls pixel processing actions such as retention, enhancement, cropping, display version generation, blurring, masking, pixelation, or downfidelity reduction. The data retention control method controls whether the original region data is stored in its original form, whether it is written to the formal image library, and whether the original region cache data is deleted. This setting can reduce the possibility of data controls such as "do not enter into the library" and "delete cache" being mistakenly executed as image processing paths, making the processing flow clearer.

[0020] 4. This invention employs a strict matching method between usage processing rules and region processing authorization information. When either prohibits a certain image processing method, prohibits display in its original form, prohibits storage in its original form, prohibits the import of original region data into the database, or requires the deletion of original region cache data, the corresponding control content is written into the prohibited image processing method, output control method, or data retention control method. This method can reduce the possibility of uncertainty in the actual image processing method due to conflicts between different processing rules.

[0021] 5. This invention binds a region source tag and a data retention control tag to the processing result of each region. After generating the purpose isolation processing result, it determines whether the image region that is prohibited from being retained, displayed or stored in its original form still corresponds to the original region data based on the region processing permission table, the region source tag and the data retention control tag. When it is found that the corresponding image region still corresponds to the original region data, at least one irreversible processing of blurring, masking, pixelation or down-fidelity is re-executed on the corresponding image region, and the purpose isolation processing result is updated, thereby reducing the risk of restricted original region data entering the corresponding business purpose image version.

[0022] 6. This invention generates a region processing record, which records the region identifier, region type, region coordinates, business purpose identifier, permitted image processing methods, prohibited image processing methods, output control methods, data retention control methods, processing priority, actual image processing method, processing result index, region source marker, data retention control marker, and cache deletion marker. Through the structured record, each image region in the purpose-isolated processing result can be mapped to the specific processing basis, actual image processing method, data source, and data retention control status, which facilitates subsequent verification and traceability. Attached Figure Description

[0023] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein: Figure 1 This is a flowchart illustrating the overall workflow of the present invention.

[0024] Figure 2 This is a sub-flowchart of the image region division in this invention.

[0025] Figure 3 This is a sub-flowchart for generating the region processing permission table of the present invention.

[0026] Figure 4 This is a flowchart of the regional differentiation processing and output verification of the present invention. Detailed Implementation

[0027] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0028] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0029] Secondly, the term "an embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention; the phrase "in an embodiment" appearing in different places in this specification does not all refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0030] Example Reference Figures 1-4 This embodiment provides a computer vision image processing method based on artificial intelligence, which is used to divide the human image to be processed into regions, and perform differentiated image processing, output control and data retention control on different image regions based on business purpose identifiers, region processing authorization information and purpose processing rule tables, thereby generating purpose isolation processing results that match the business purpose.

[0031] The method is executed by a computer device with a processor and memory; the computer device can be a server, edge computing terminal, mobile terminal, or image review device. The computer device pre-stores image processing programs, preset parameter tables, usage processing rule tables, region processing method libraries, and computer vision model sets.

[0032] The computer vision model set includes a pre-trained human semantic segmentation model, a facial landmark detection model, a head and shoulder contour segmentation model, and an accessory detection model. The human semantic segmentation model outputs the subject mask and non-subject masks; the facial landmark detection model outputs facial landmarks; the head and shoulder contour segmentation model outputs the head contour, hairstyle contour, and shoulder and neck contour; and the accessory detection model outputs accessory categories, detection boxes, accessory confidence scores, and accessory masks.

[0033] The purpose-based processing rule table records the default processing control methods for each image region under different business purposes. The purpose-based processing rule table includes at least the business purpose identifier, region type, default allowed image processing methods, default prohibited image processing methods, default output control methods, default data retention control methods, and default processing priority.

[0034] Image processing methods include at least one of the following: preservation, enhancement, cropping, display version generation, blurring, masking, pixelation, and fidelity reduction. Output control methods include at least one of the following: allow display, prohibit display in original form, allow output, and prohibit output. Data retention control methods include at least one of the following: allow storage, prohibit storage in original form, prohibit original region data from being stored, and delete original region cached data. The region processing method library is used to record the processing instructions or control instructions corresponding to each image processing method, output control method, and data retention control method.

[0035] The region processing authorization information is generated by the business system based on the business purpose identifier, business scenario configuration, and user authorization result, and is stored in correspondence with the image processing task number. The user authorization result is the data generated after the user confirms the image region processing range on the authorization page. The region processing authorization information includes at least the image processing task number, business purpose identifier, region type, authorized image processing method, prohibited image processing method, permitted output control method, prohibited output control method, permitted data retention control method, and prohibited data retention control method.

[0036] S1. Receive the image of the person to be processed, the image processing task number, the business purpose identifier, and the area processing authorization information; Reference Figure 1 As shown, this step is used to obtain the raw input data and business control information required for image processing, providing a data foundation for subsequent image region segmentation and permission determination. The image processing task is initiated by the business system. The computer device needs to receive the portrait image to be processed and the corresponding task number, and query the business system based on the task number to obtain the business purpose identifier and region processing authorization information. Finally, all input information is associated to form a structured image processing input data packet to ensure that subsequent steps can accurately identify the business scenario and authorization boundaries to which the current processing task belongs.

[0037] S11. Receive the portrait image to be processed and the image processing task number; Upon receiving an image processing request, the business system creates an image processing task and generates an image processing task number. The computer equipment receives the portrait image to be processed uploaded by the image acquisition terminal or the business system, and simultaneously receives the image processing task number issued by the business system. The image processing task number is used to associate intermediate data, output data, business purpose identifier, and area processing authorization information during this image processing process.

[0038] S12, Read the business purpose identifier; The computer device sends a purpose query request to the business system based on the image processing task number, and receives and reads the business purpose identifier returned by the business system. The business purpose identifier is generated by the business system when creating the image processing task based on the business page entry, business process node, or business submission type, and is stored in correspondence with the image processing task number.

[0039] The business purpose identifier includes at least one of document submission, headshot display, background filing, manual review, and material archiving, and is used to determine the target business purpose version that the portrait image to be processed needs to be generated subsequently.

[0040] S13, Read the region processing authorization information; The computer equipment sends an authorization information query request to the business system based on the image processing task number, and receives and reads the area processing authorization information returned by the business system.

[0041] The region processing authorization information is used to limit whether image processing such as enhancement, cropping, display version generation, blurring, masking, pixelation, or down-fidelity reduction is allowed in the face region, hairstyle region, shoulder and neck region, background region, and accessory region, as well as whether it is allowed to be displayed in its original form, stored in its original form, written to the official image library, or have its original region cache data deleted.

[0042] S14. Generate image processing input data packets; The computer equipment associates the image to be processed, the image processing task number, the business purpose identifier, and the area processing authorization information into the image buffer, forming an image processing input data packet. This image processing input data packet serves as input data for subsequent image compliance verification, image region division, and the generation of the region processing permission table.

[0043] S2. Perform image compliance verification on the human portrait image to be processed; After constructing the image processing input data package, this step performs multi-dimensional compliance checks on the portrait images to be processed, including format, size, resolution, and file size, to prevent non-compliant images from entering the subsequent model inference stage. By pre-screening the basic image attributes, compliance checks can identify and exclude images that do not meet processing standards at an early stage, reducing unnecessary computational resource consumption and providing consistent and reliable input images for subsequent image region segmentation steps.

[0044] S21. Read the image compliance parameters; The computer device reads the image compliance parameters from the preset parameter table. The image compliance parameters include the preset image format, preset pixel size range, preset resolution range, and preset file size range.

[0045] S22. Perform image compliance verification; The computer equipment performs image format verification, pixel size verification, resolution verification, and file size verification on the portrait image to be processed based on image compliance parameters. Specifically, image format verification determines whether the file format of the portrait image to be processed belongs to a preset image format; pixel size verification determines whether the pixel width and pixel height of the portrait image to be processed are within a preset pixel size range; resolution verification determines whether the resolution of the portrait image to be processed is within a preset resolution range; and file size verification determines whether the data size of the portrait image to be processed is within a preset file size range.

[0046] S23. Process the image compliance verification results; When the image compliance verification result is passed, the computer device writes the image compliance verification result into the image processing input data packet and proceeds to step S3. When the image compliance verification result is failed, the computer device generates an image compliance exception record and proceeds to step S8 to output a re-upload prompt.

[0047] S3. Use computer vision models to segment the image regions in the portrait image to be processed. After the image compliance verification is passed, refer to Figure 2 As shown, this step inputs the image of the person to be processed into multiple pre-trained artificial intelligence visual models. Through collaborative reasoning of multiple models, such as portrait semantic segmentation, facial key point detection, head and shoulder contour segmentation, and accessory detection, a refined decomposition of the main body region of the portrait is achieved. The core purpose of this step is to divide the entire image into multiple image regions with independent semantic meanings, such as the face region, hairstyle region, shoulder and neck region, background region, and accessory region. For each region, a corresponding region identifier, region type, region coordinates, and region mask are generated, so that subsequent image processing can be performed within a specific pixel range, avoiding the inter-regional errors caused by uniform processing of the entire image.

[0048] S31. Perform portrait subject segmentation; The computer device reads the portrait image to be processed from the image processing input data packet, inputs the portrait image to be processed into a pre-trained portrait semantic segmentation model, and outputs a portrait subject mask and a non-portrait subject mask. The computer device determines the portrait subject region based on the portrait subject mask and determines the non-portrait subject region based on the non-portrait subject mask.

[0049] S32, further subdivide the main portrait area; The computer device inputs the image of the subject into a facial landmark detection model to obtain facial landmarks; it also inputs the image into a head and shoulder contour segmentation model to obtain head contour, hairstyle contour, and shoulder and neck contour. Within the main subject area of ​​the image, the computer device divides the image into a face region, a hairstyle region, and a shoulder and neck region based on the facial landmarks, head contour, hairstyle contour, and shoulder and neck contour, and further divides the image into a background region within the non-subject area.

[0050] S33, Area for detecting attachments; The computer device inputs the human image to be processed into the accessory detection model and outputs accessory categories, detection boxes, accessory confidence scores, and accessory masks. When the accessory confidence score reaches a preset confidence threshold, the corresponding accessory is classified into one or more accessory regions; when the accessory confidence score does not reach the preset confidence threshold, the corresponding candidate region is not written into the accessory region.

[0051] Accessories include glasses, hats, masks, headphones, jewelry, ID badges, or other personal items.

[0052] S34. Generate image region segmentation results; The computer device generates region identifiers, region types, region coordinates, and region masks for face regions, hairstyle regions, shoulder and neck regions, background regions, and accessory regions, respectively, forming the image region segmentation results. The region coordinates are determined by the outer boundary of the corresponding region mask; the region mask is determined by the set of pixel positions of the corresponding image region; when multiple regions of the same type exist, different region identifiers are generated for each region.

[0053] For image regions that are adjacent or partially overlapping, the computer device determines the region assignment of overlapping pixels according to a preset region priority. The preset region priority is set from high to low in the order of accessory regions, face regions, hairstyle regions, shoulder and neck regions, and background regions. As a result, accessory pixels can be preferentially separated from face regions, hairstyle regions, or shoulder and neck regions, reducing the possibility of accessory pixels being mistakenly included in human body regions.

[0054] S35, Binding area and input data; The computer device binds the image region division results with the image processing task number, business purpose identifier, and region processing authorization information, and then proceeds to step S4.

[0055] S4. Generate the regional processing permission table; After obtaining the spatial location and mask information of each image region, refer to Figure 3As shown, this step retrieves the default processing rules for each region type from the purpose processing rule table based on the business purpose identifier, and then merges and matches them with the region processing authorization information to generate a corresponding region processing permission table for each image region. This permission table defines the allowed and prohibited image processing methods, output control methods, data retention control methods, and processing priorities for each region, serving as the direct basis for subsequent region-specific processing. The permission generation process adopts a "strict matching" principle; when there is a conflict between the purpose processing rule and the region processing authorization information, the more stringent restriction prevails, thereby ensuring that the processing permission boundaries are clear and controllable.

[0056] S41, Read usage processing rules; The computer equipment reads the usage processing rules corresponding to the usage identifier and region type from the usage processing rule table based on the usage identifier and the region type in the image region division results. The usage processing rules include the default allowed image processing methods, the default prohibited image processing methods, the default output control methods, the default data retention control methods, and the default processing priority.

[0057] S42. Merge regional processing authorization information; The computer device matches the usage processing rules with the region processing authorization information: it matches the region type in the image region division result with the region type in the region processing authorization information, and compares the default processing control method with the authorized processing control method under the same region type.

[0058] When either the purpose processing rule or the region processing authorization information prohibits a certain image processing method, the image processing method is written into the prohibited image processing method in the region processing permission table; when the purpose processing rule allows a certain image processing method and the region processing authorization information does not prohibit the image processing method, the image processing method is written into the allowed image processing method in the region processing permission table.

[0059] When either the usage processing rules or the region processing authorization information prohibits displaying data in its original form, prohibits storing data in its original form, prohibits importing original region data into the database, or requires the deletion of original region cache data, the corresponding content will be written into the output control method or data retention control method. Therefore, when there is a conflict between the region processing authorization information and the usage processing rules, the one with the stricter restriction will be used as the basis for processing permissions.

[0060] S43. Create a regional processing permission table; The computer equipment summarizes the region identifier, region type, region coordinates, region mask, allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priorities to form a region processing permission table.

[0061] When both image processing and data retention control methods exist in the same image area, the data retention control method is given priority to restrict the storage, entry into the database, or cache deletion of the original area data, and the actual image processing method is determined under the condition that the data retention control method is satisfied; when both protective image processing and enhancement image processing methods exist in the same image area, blurring, masking, pixelation, or down-fidelity reduction is given priority to be determined as the actual image processing method.

[0062] S44. Verify the area processing permission table; The computer device determines whether there are image regions without configured processing permissions in the region processing permission table. If there are, it configures a default processing control method for the image region based on the region type, business purpose identifier, or risk level. The default processing control method includes at least one of the following: masking, pixelation, down-fidelity, prohibiting display in the original form, or prohibiting the original region data from being entered into the database, and updates the region processing permission table. If there are no such regions, the region processing permission table is directly entered into step S5.

[0063] S5. Perform regional differential processing based on the regional processing permission table; After the regional processing permission table is generated, refer to Figure 4 The flowchart shown illustrates the regional differentiation processing and output verification process. This step performs differentiated pixel-level processing operations on different image regions based on their respective processing permissions. Processing methods include preservation, enhancement, cropping, display version generation, and irreversible protective processing such as blurring, masking, pixelation, and fidelity reduction. Simultaneously, this step also controls the entry, storage, and cache deletion of original region data according to data retention control methods, and binds the processing result of each region to its region source tag and data retention control tag, forming a traceable set of region processing results.

[0064] S51. Determine the actual image processing method and data retention control method for each area; The computer device traverses the area processing permission table, and based on the allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priorities corresponding to each image area, reads the corresponding processing instructions or control instructions from the area processing method library to determine the actual image processing method and data retention control method for each image area.

[0065] When an image processing method appears in both permitted and prohibited image processing methods, the prohibited image processing method shall prevail. When an image region is prohibited from being displayed in its original form, stored in its original form, or having its original region data imported into the database, the reserved region shall not be determined as the actual image processing method for that image region, or the processing result of the reserved region shall not be used as the output region for the corresponding business application version.

[0066] S52. Process the areas that are to be retained in the actual image processing method; For image regions that are preserved in their original form and are not prohibited from being preserved, displayed, or stored in their original form, the computer device copies the original region data of the image region according to the corresponding region coordinates and region mask, generates the preserved region processing result, and writes the region source mark as the original region data.

[0067] S53. Process areas where the actual image processing method is enhancement; For image regions where the actual image processing method is enhancement and no protective image processing is required, the computer device performs brightness equalization, noise suppression, edge preservation, or color correction processing based on the corresponding region coordinates and region mask to generate the enhanced region processing result, and writes the region source mark as enhanced region data.

[0068] Among them, brightness equalization is used to adjust the brightness distribution of the corresponding image area to the target brightness range; noise suppression is used to reduce random noise in the corresponding image area; edge preservation is used to preserve the area boundary and main contour when reducing noise; and color correction is used to adjust the color shift of the corresponding image area.

[0069] S54. Process the area generated by the actual image processing method of cropping or displaying the version; For image regions whose actual image processing method is cropping, or image regions whose output control method allows the generation of display versions, the computer device performs boundary cropping, size adaptation, or display version generation processing based on the corresponding region coordinates and region mask, generating cropped region processing results or display version region processing results.

[0070] Specifically, boundary cropping is used to extract the corresponding image region according to the region coordinates; size adaptation is used to adjust the corresponding image region to the size required for the target business application; and display version generation processing is used to generate a region image with the corresponding resolution or display ratio according to the display purpose. For the cropped region processing result, the region source tag is written into the cropped region data; for the display version region processing result, the region source tag is written into the display version region data.

[0071] S55. Process areas where the actual image processing method is protective image processing; For image regions that are actually blurred, masked, pixelated, or de-fidelity processed, the computer device performs irreversible processing based on the corresponding region coordinates and region mask, generates irreversible region processing results, and writes the region source marker into the irreversible processing region data.

[0072] Among them, blurring is used to reduce the detail clarity of the corresponding image area; masking is used to cover the corresponding image area with a preset color, pattern or mask; pixelation is used to divide the corresponding image area into pixel blocks and replace the original pixels with the average or approximate value of the pixel blocks; and downfidelity is used to reduce the resolution, texture detail or compression quality of the corresponding image area.

[0073] S56. Implement data retention control; Based on the data retention control method corresponding to each image region determined in step S51, a corresponding data retention control marker is generated for each image region. The data retention control marker is used to indicate the data retention control method applicable to the corresponding image region; when two or more data retention control methods are applicable to the same image region, the data retention control marker records multiple corresponding data retention control items.

[0074] For image regions where data retention control is set to allow storage, the corresponding region processing results are stored according to the actual image processing method, provided that the prohibited image processing method and output control method for that image region are satisfied, and the data retention control flag is written as allowed storage.

[0075] For image regions whose data retention control method prohibits storage in their original form, the original region data corresponding to the image region is not stored in its original form. When the business application requires storing the corresponding region results, only the irreversible region processing results, region processing tags, or processing result indexes after blurring, masking, pixelation, or down-fidelity processing are stored, and the data retention control tag is written as prohibiting storage in its original form.

[0076] For image regions where the data retention control method prohibits the entry of original region data into the database, the computer equipment will not write the original region data corresponding to that image region into the official image database. When business purposes require the retention of the corresponding region results, only the irreversibly processed region data, region processing tags, or processing result indexes will be written into the result database or record database, and the data retention control tag will be set to prohibit the entry of original region data into the database.

[0077] For image regions where the data retention control method is to delete the original region cache data, the data retention control flag is written as "delete original region cache data". After generating the purpose isolation processing result, a cache to be deleted flag is generated and written into the region processing record. After the corresponding original region cache data is deleted from the image cache area to be processed, the cache to be deleted flag is updated to the cache deletion completed flag.

[0078] The data retention control flag is used to record the data retention control method applicable to the corresponding image area. The cache to be deleted flag and the cache deletion completed flag are used to record the execution status of the data retention control method of deleting cache data in the original area.

[0079] S57. Form a set of regional processing results; The computer device will summarize the generated data from the retained region processing results, enhanced region processing results, cropped region processing results, displayed version region processing results, and irreversible region processing results to form a set of region processing results.

[0080] Each region processing result in the region processing result set is bound to a corresponding region identifier, region type, actual image processing method, processing result index, region source marker, and data retention control marker. The region source marker is used to distinguish whether the corresponding region processing result originates from the original region data, enhanced region data, cropped region data, display version region data, or irreversibly processed region data.

[0081] S6. Generate the results of the application isolation process; After regional differentiation processing is completed, this step determines the target business purpose version based on the business purpose identifier and reads the processing results of each region corresponding to that business purpose from the regional processing result set. According to the regional coordinates and regional masks of each region, the regional processing results from different sources are recombined into a complete purpose-isolated processing image. Purpose-isolated processing results corresponding to different business purposes are generated and stored separately, enabling the same portrait image to output mutually isolated image processing versions in different business scenarios, avoiding interference between image processing methods for one purpose and versions for other purposes.

[0082] S61. Determine the version for the target business purpose; The computer equipment extracts the service purpose identifier from the image processing input data packet and determines the target service purpose version to be generated based on the service purpose identifier. When the same portrait image to be processed corresponds to multiple service purpose identifiers, multiple target service purpose versions are determined separately.

[0083] S62. Read the processing results for the corresponding region; The computer equipment reads the region processing result corresponding to the target business purpose version from the region processing result set according to the preset purpose region mapping relationship. The purpose region mapping relationship is used to record the type of region processing result that should be adopted for each region type under different business purpose versions.

[0084] For example, for document submission purposes, the face area can use either the preserved area processing result or the enhanced area processing result, while the background area can use the irreversible area processing result; for avatar display purposes, the face area can use the display version area processing result, while the accessory area can use either the preserved area processing result or the irreversible area processing result according to the area processing permission table; for backend archiving purposes, areas where original area data is prohibited from being entered into the database only use the irreversible area processing result or the processing result index.

[0085] S63, Results of Reorganization Area Processing; The computer device reconstructs the image by processing the results of each region according to the region coordinates and region masks. During image reconstruction, the pixel position of each region's processing result in the output image is determined based on the region masks; when multiple region masks overlap, the coverage relationship of the overlapping pixels is determined according to a preset region priority.

[0086] S64. Generate the results of the application isolation process; The computer equipment generates usage isolation processing results corresponding to the business usage identifier. The usage isolation processing results corresponding to different business uses are generated, stored, or output separately, so that the same portrait image to be processed forms mutually isolated image processing versions under different business uses.

[0087] S7. Verify the isolation processing results. After the purpose isolation processing results are generated, refer to Figure 4 The latter half of the process, as shown, involves output verification of the generated usage isolation processing results. The core purpose of this verification is to determine whether image regions prohibited from being retained, displayed, or stored in their original form by the region processing permission table still correspond to the original region data. If the verification finds that restricted original region data has entered the output image, irreversible processing is re-executed on that region, and the usage isolation processing results are updated. This forms a final data protection barrier before image output, reducing the risk of restricted region original data being leaked with the output version.

[0088] S71, Read the region source marker and data retention control marker; The computer equipment reads the region source marker and data retention control marker of each region in the purpose isolation processing result, and reads the prohibited image processing methods, output control methods and data retention control methods of the corresponding region in the region processing permission table.

[0089] S72. Determine if there is any restricted original region data; The computer device determines, based on the region processing permission table, whether an image region that is prohibited from being retained, displayed, or stored in its original form still corresponds to the original region data. If the region source marker indicates that the corresponding region originates from the original region data, and the region is prohibited from being retained, displayed, or stored in its original form in the region processing permission table, then it is determined that the purpose isolation processing result contains restricted original region data.

[0090] S73, Update the results of the isolation treatment for the updated purpose; When restricted original region data is detected, the computer device re-performs blurring, masking, pixelation, or down-fidelity processing on the corresponding image region, generating a new irreversible region processing result. The purpose isolation processing result is then updated based on this new irreversible region processing result. After the update, the computer device reads the region source marker again and performs verification.

[0091] If the verification passes again, proceed to step S8; if the verification fails again, generate a record of the isolated exception and proceed to step S8 to output the exception message.

[0092] S8. Generate region processing records and output image processing results; After the purpose isolation processing result is verified, this step summarizes the key data from the entire image processing workflow to generate a structured region processing record. This record covers information such as the region identifier, region type, spatial coordinates, processing permission configuration, actual image processing method, region source marker, data retention control marker, and cache deletion marker for each image region. The region processing record is then stored in association with the purpose isolation processing result. Finally, the computer equipment outputs the purpose isolation processing result and the corresponding region processing record to the business system, providing a complete basis for subsequent image review, issue tracing, and compliance auditing.

[0093] S81, Generate region processing records; The computer equipment reads the image processing task number, business purpose identifier, image region division result, region processing permission table, region processing result set, and purpose isolation processing result, and generates a region processing record.

[0094] The region processing record includes region identifier, region type, region coordinates, business purpose identifier, permitted image processing methods, prohibited image processing methods, output control methods, data retention control methods, processing priority, actual image processing method, processing result index, region source marker, data retention control marker, and cache deletion marker.

[0095] S82, Results of isolation processing for related uses; The computer device associates and stores the area processing records with the usage isolation processing results, and each processing area in the usage isolation processing results can be mapped to an area processing record.

[0096] S83, Output the results of the isolation processing and the regional processing records; When the purpose isolation processing result passes verification, the computer device outputs the purpose isolation processing result to the business system and outputs or stores the area processing record. The purpose isolation processing result is used for image submission, display, review, or archiving under the corresponding business purpose, and the area processing record is used for subsequent traceability of the area processing method, area source, and cache deletion status in the purpose isolation processing result.

[0097] When an image compliance error record is generated, the computer device outputs a prompt to re-upload; when a record for purpose isolation processing error is generated, the computer device outputs a processing error prompt.

[0098] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A computer vision image processing method based on artificial intelligence, characterized in that, This includes performing the following steps in sequence: Receive the portrait image to be processed and the image processing task number; The business purpose identifier and regional processing authorization information are obtained based on the image processing task number, and the portrait image to be processed, the image processing task number, the business purpose identifier, and the regional processing authorization information are associated to form an image processing input data packet; The image of the person to be processed is input into a pre-trained computer vision model group to perform image region segmentation and generate image region segmentation results. The image region segmentation results include the region identifier, region type, region coordinates and region mask of each image region. Among them, for image regions that are adjacent or partially overlap, the ownership of overlapping pixels is determined according to a preset region ownership rule. Based on the business purpose identifier, area processing authorization information, purpose processing rule table, and area type in the image area division result, an area processing permission table corresponding to each image area is generated. The area processing permission table includes area identifier, area type, area coordinates, area mask, allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priority. Based on the region processing permission table and image region division results, region-differentiated processing is performed on each image region based on region coordinates and the scope of action limited by region masks, generating a set of region processing results. Each region processing result in the set is bound to a corresponding region identifier, actual image processing method, region source marker, and data retention control marker. The region source marker is used to identify the data source type of the corresponding region processing result, and the data retention control marker is used to identify the data retention control method applicable to the corresponding image region. Based on the business purpose identifier and the preset purpose area mapping relationship, the corresponding area processing result is read from the area processing result set, and the purpose isolation processing result is generated by recombining the area coordinates and area mask. Based on the region processing permission table, region source marker, and data retention control marker, the purpose isolation processing result is output verified to determine whether the image region that is prohibited from being retained, displayed, or stored in its original form still corresponds to the original region data. When the determination result is yes, at least one irreversible processing of blurring, masking, pixelation, or down-fidelity is re-executed on the corresponding image region, and the purpose isolation processing result is updated based on the re-processed region processing result. When the purpose isolation processing result does not contain any image regions that are prohibited from being retained, displayed, or stored in their original form but still correspond to the original region data, a region processing record associated with the purpose isolation processing result is generated.

2. The computer vision image processing method based on artificial intelligence as described in claim 1, characterized in that: The business purpose identifier is generated by the business system when creating an image processing task based on the business page entry, business process node, or business submission type, and is stored in correspondence with the image processing task number. The regional processing authorization information is generated by the business system based on the business purpose identifier, business scenario configuration and user authorization result, and is stored in correspondence with the image processing task number. The region processing authorization information includes at least the region type, authorized image processing methods, prohibited image processing methods, permitted output control methods, prohibited output control methods, permitted data retention control methods, and prohibited data retention control methods.

3. The computer vision image processing method based on artificial intelligence as described in claim 2, characterized in that: Before performing image region segmentation, the following steps are also included: reading image compliance parameters from a preset parameter table, including preset image format, preset pixel size range, preset resolution range, and preset file size range; performing image format verification, pixel size verification, resolution verification, and file size verification on the portrait image to be processed according to the image compliance parameters; performing image region segmentation when the image compliance verification result is passed; and generating an image compliance exception record when the image compliance verification result is failed.

4. The computer vision image processing method based on artificial intelligence as described in claim 3, characterized in that: The computer vision model set includes a pre-trained portrait semantic segmentation model, a facial landmark detection model, a head and shoulder contour segmentation model, and an accessory detection model. The image region segmentation steps include: The human image to be processed is input into the human semantic segmentation model, which outputs a human subject mask and a non-human subject mask. The human subject region is determined based on the human subject mask, and the non-human subject region is determined based on the non-human subject mask. Within the main human figure area, the face region, hairstyle region, and shoulder and neck region are divided based on the facial key points output by the facial key point detection model and the head contour, hairstyle contour, and shoulder and neck contour output by the head and shoulder contour segmentation model. Define the background area within the non-human subject area; The model inputs the human image to be processed into the accessory detection model and outputs the accessory category, detection box, accessory confidence score, and accessory mask. When the accessory confidence score reaches the preset confidence threshold, the corresponding accessory is divided into one or more accessory regions. Generate corresponding region coordinates based on the region mask of each image region, and generate region identifiers and region types for each image region; For image regions that are adjacent to or partially overlap each other, the region assignment of overlapping pixels is determined according to the preset region priority. The preset region priority is set from high to low in the order of accessory region, face region, hairstyle region, shoulder and neck region and background region.

5. The computer vision image processing method based on artificial intelligence as described in claim 4, characterized in that: The steps for generating the region processing permission table include: Based on the business purpose identifier and region type, the corresponding purpose processing rules are read from the purpose processing rule table. The purpose processing rules include the default allowed image processing method, the default prohibited image processing method, the default output control method, the default data retention control method, and the default processing priority. The usage processing rules are matched with the regional processing authorization information; When either the usage processing rule or the area processing authorization information prohibits a certain image processing method, the image processing method is written into the prohibited image processing methods; When the application processing rule allows a certain image processing method and the region processing authorization information does not prohibit the image processing method, the image processing method is written into the allowed image processing methods. When any of the aforementioned usage processing rules and regional processing authorization information prohibits display in its original form, prohibits the entry of original regional data into the database, prohibits storage in its original form, or requires the deletion of original regional cache data, the corresponding control content will be written into the output control method or data retention control method. The region identifier, region type, region coordinates, region mask, allowed image processing methods, prohibited image processing methods, output control methods, data retention control methods, and processing priority are summarized to form a region processing permission table.

6. The computer vision image processing method based on artificial intelligence as described in claim 5, characterized in that: The image processing method includes at least one of the following: preservation, enhancement, cropping, display version generation, blurring, masking, pixelation, and defidelity reduction; The output control method includes at least one of allowing display, prohibiting display in its original form, allowing output, and prohibiting output. The data retention control method includes at least one of the following: allowing storage, prohibiting storage in its original form, prohibiting the entry of original area data into the database, and deleting original area cache data. When both image processing and data retention control methods exist in the same image area, the data retention control method is executed first to restrict the storage, entry into the database, or cache deletion of the original area data, and the actual image processing method is determined under the condition that the data retention control method is satisfied. When the allowed image processing methods for the same image region include enhancement as well as at least one of blurring, masking, pixelation, or defidelity, the actual image processing method shall be determined first from blurring, masking, pixelation, or defidelity. When there is an image region in the region processing permission table that has no processing permission configured, a default processing control method is configured for the image region according to the region type and business purpose identifier. The default processing control method includes at least one of the following: masking, pixelation, down-fidelity, prohibiting display in the original form, or prohibiting the original region data from being entered into the database.

7. The computer vision image processing method based on artificial intelligence as described in claim 6, characterized in that: The regional differentiation processing includes: For image regions that are preserved in their original form and are not prohibited from being preserved, displayed or stored in their original form, the original region data of the image region is copied according to the corresponding region coordinates and region mask to generate the preserved region processing result. For image regions where the actual image processing method is enhancement, brightness equalization, noise suppression, edge preservation, or color correction are performed based on the corresponding region coordinates and region mask to generate the enhancement region processing result. For image regions generated by actual image processing methods such as cropping or display, perform boundary cropping, size adaptation, or display version generation processing based on the corresponding region coordinates and region mask to generate cropped region processing results or display version region processing results. For image regions that are actually blurred, masked, pixelated, or de-fidelity processed, irreversible processing is performed based on the corresponding region coordinates and region mask to generate irreversible region processing results.

8. The computer vision image processing method based on artificial intelligence as described in claim 7, characterized in that: For image regions where the data retention control method prohibits the entry of original regional data into the database, the original regional data corresponding to the image region will not be written into the formal image database; when the business needs to retain the corresponding regional results, only the irreversibly processed regional data or the processing result index will be written into the result database or record database. For image regions where the data retention control method is to delete the original region cache data, a cache mark to be deleted is generated after the purpose isolation processing result is generated, and the cache mark to be deleted is associated with the region identifier of the corresponding image region; When generating a region processing record, the cache to be deleted marker is written into the region processing record; after the corresponding original region cache data is deleted, the cache to be deleted marker in the region processing record is updated to a cache deletion completed marker.

9. The computer vision image processing method based on artificial intelligence as described in claim 8, characterized in that: Each region processing result in the region processing result set is bound to the corresponding region identifier, region type, actual image processing method, processing result index, region source marker, and data retention control marker; The region source marker is used to distinguish whether the corresponding region processing result comes from the original region data, enhanced region data, cropped region data, display version region data, or irreversibly processed region data. When generating the purpose isolation processing result, the target business purpose version is determined according to the business purpose identifier, the region processing result corresponding to the target business purpose version is read according to the preset purpose region mapping relationship, and the image is reconstructed according to the region coordinates and region mask. When the region masks corresponding to multiple image regions overlap, the coverage relationship of overlapping pixels is determined according to the preset region priority.

10. The computer vision image processing method based on artificial intelligence as described in claim 9, characterized in that: After generating the purpose isolation processing result, the region source mark and data retention control mark of each region in the purpose isolation processing result are read, and the region processing permission table is used to determine whether the image region that is prohibited from being retained, displayed or stored in its original form still corresponds to the original region data. When the initial judgment result is yes, at least one irreversible processing of blurring, masking, pixelation or downfidelity reduction is re-executed on the corresponding image region, and the purpose isolation processing result is updated based on the re-processed region processing result. The region source tag and data retention control tag of the corresponding image region in the updated purpose isolation processing result are read, and the region processing permission table is used to determine again whether the corresponding image region still corresponds to the original region data. If the result is yes again, generate an exception record for isolation processing. If the initial judgment result is negative, or if the second judgment result is negative, a region processing record is generated. The region processing record includes region identifier, region type, region coordinates, business purpose identifier, allowed image processing method, prohibited image processing method, output control method, data retention control method, processing priority, actual image processing method, processing result index, region source marker, data retention control marker, and cache deletion marker.

Images