A video and picture desensitization method and device, electronic equipment and storage medium

Abstract

The application provides a video and picture desensitization method and device, electronic equipment and storage medium, the method comprises the following steps: acquiring the image sensitive area of a video or a picture; performing slice segmentation on the image sensitive area; using different motion blurring methods to blur each part of the slice segmentation; merging the processed parts to obtain a composite image of the desensitization area, and then synthesizing the image to output the final result image. The application uses the motion blurring method to desensitize the sensitive area, and completes the desensitization under the premise of maintaining the overall layout and beauty of the picture; and according to the reversibility of the motion blurring, the sensitive area is sliced, different direction motion blurring algorithms are used for sensitive area desensitization according to different slice areas, so that the desensitized picture cannot be restored, thereby achieving the dual effects of picture desensitization and maintaining the layout and beauty of the picture.

Description

TECHNICAL FIELD

[0001] The present application relates to the desensitization technology of video and image, and in particular to a method for desensitizing sensitive areas (such as faces or license plate numbers, etc.) in videos and pictures taken by vehicles. BACKGROUND

[0002] As a common means of transportation, automobiles are rapidly developing towards electrification and intelligentization. Nowadays, people not only pay attention to the mechanical vehicles themselves, but also to the friendly interaction between people and vehicles and the connection. The rise of software-defined architecture or software-defined automobile technology brings more possibilities and imagination space for the intelligentization of vehicles and the experience of people-vehicle interaction.

[0003] Due to the rapid popularization and application of networked vehicles, vehicles are becoming more and more intelligent, and the resolution of cameras mounted on vehicles is also getting higher and higher. Vehicle cameras can assist users in driving and monitor the surroundings of vehicles. In the intelligent vehicle cloud scenario, during parking, users can remotely take pictures of the surroundings of the vehicle through a mobile phone app, and then the vehicle uploads the taken pictures to the cloud. Users can view the pictures taken by the vehicle through the app, which is so-called remote photography.

[0004] According to the requirements of the regulation “Requirements and Methods for Desensitization of Video and Image Transmitted by Automobiles”, the video and image collected by vehicles need to be desensitized. Desensitization refers to a data processing process that eliminates sensitive information in original environmental data on a vehicle-side data processing device, so that the information subject cannot be identified or associated, and the processed information cannot be restored, while retaining the data features or contents required by the target environmental business.

[0005] Currently, the desensitization method used for vehicle video and picture desensitization is basically pixel erasing or color block replacing in sensitive areas. Using color block replacement and pixel erasing will destroy the overall visual effect of the picture. Compared with other desensitization methods, motion blurring processing can make the sensitive areas such as faces or license plate numbers in the processed image look more natural and beautiful. Therefore, motion blurring processing technology is also used to realize video and picture desensitization. Motion blurring is caused by many reasons that cause image degradation or blurring. If the blurring is caused by the relative motion between the camera and the photographed object during shooting, it is called motion blurring. Common motion blurring algorithms include line integral convolution (LIC) motion blurring and motion blurring with a standard mapping function. Because of the commonality of motion blurring effect, research on the recovery processing of motion-blurred pictures is a common research direction, and some motion blurring recovery algorithms such as Wiener filter are developed to restore the blurring effect. However, when applied to image desensitization in image processing, it is necessary to consider that the desensitized pictures are not recoverable.

[0006] For example, the prior art document CN115049540A discloses an image desensitization method, device, electronic equipment and storage medium, wherein the desensitization method comprises: receiving a to-be-desensitized image; receiving a type of a target region to be obtained; determining a first algorithm unit in a first algorithm pool according to the type of the target region to be obtained; inputting the to-be-desensitized image into the first algorithm unit, so that the first algorithm unit obtains the target region according to the to-be-desensitized image; determining a second algorithm unit in a second algorithm pool; inputting the target region into the second algorithm unit, so that the second algorithm unit desensitizes the target region to obtain a desensitized target region; and embedding the desensitized target region into the to-be-desensitized image to obtain a desensitized image. By selecting a suitable algorithm unit according to the type of the target region, the target region is accurately segmented and desensitized. The desensitization method comprises: desensitizing the target region by using one or more of affine deformation, motion blur, opening operation, dilation, and mosaic algorithm to obtain the desensitized target region. As described above, in the desensitization process, although one or more of affine deformation, motion blur, opening operation, dilation, and mosaic algorithm are proposed to desensitize the target region, the desensitized region is desensitized as a whole, and the processed picture still has a high reversibility.

[0007] Therefore, the prior art does not have an effective way to achieve the dual effects of picture desensitization and maintaining the layout and beauty of the picture, and to realize the unrecoverable desensitized picture. SUMMARY

[0008] In view of the deficiencies of the prior art, the present application provides a video and picture desensitization method, device, electronic equipment and storage medium, which desensitizes sensitive regions (such as faces or license plate numbers) in the picture by using an irreversible motion blur processing method, and makes the desensitized region unrecoverable.

[0009] The technical solutions of the present application are as follows:

[0010] In a first aspect, the present application provides a video and picture desensitization method, which comprises the following processes:

[0011] First, the image sensitive region of the video or picture is obtained;

[0012] Then, the image sensitive region is sliced and segmented;

[0013] Next, different motion blurring methods are used to blur the segmented parts in different directions and to different degrees;

[0014] The processed parts are combined to obtain a composite image of the desensitized region, and then combined back into the original video image frame or the original image to output the final desensitized result image.

[0015] The blurring process applied in this invention involves slicing different parts or regions of the sensitive area, using different blurring directions for different slices, and employing a random degree of blurring. Therefore, the blurred image is unrecoverable.

[0016] Furthermore, in this desensitization method, the specific process of blurring each part of the slice using different motion blurring methods involves changing the direction and displacement of the vector fields of each part to give them random characteristics. This achieves blurring of the sensitive region at different degrees in different directions. At least two parts have different directions and displacements of their vector fields; ideally, each part should have different directions and displacements to achieve better irreversibility.

[0017] Furthermore, in this desensitization method, when performing motion fuzzification, different directions are used for fuzzification operations, and the operation vector directions are randomly generated.

[0018] Furthermore, in this desensitization method, the motion blur algorithm for each segment is randomly selected, either the same or different algorithms.

[0019] Furthermore, in this desensitization method, the motion blur algorithm is selected from the line integral convolution (LIC) motion blur algorithm and the normalized mapping function motion blur algorithm.

[0020] Furthermore, in this desensitization method, the slice segmentation adopts a random cutting method, and the cutting direction, size, shape and number are randomly determined.

[0021] Furthermore, the slice segmentation specifically includes:

[0022] Slicing is based on specific values, where the specific values ​​are grayscale values, average values, or variances.

[0023] Slicing based on regions involves dividing an image into several regions based on pixels, and then further dividing the image into several parts based on these regions.

[0024] Slicing is performed using the edge as the dimension. By detecting contours in the image, the edge regions of a portion of the image are located.

[0025] Furthermore, the region-based slicing segmentation employs any one of the following methods or a combination thereof: horizontal and vertical cutting, rotational cutting around the central region, or circular cutting centered on a circle.

[0026] In a second aspect, the present invention also provides an irreversible motion-blurred video and image desensitization device, the desensitization device comprising the following functional modules:

[0027] The sensitive area acquisition module is used to acquire the sensitive areas of a video or image.

[0028] The slice segmentation module is used to slice and segment sensitive regions of an image;

[0029] The motion blur processing module is used to blur each part of the slice using different motion blur methods.

[0030] The image compositing module is used to merge the processed parts to obtain a composite image of the desensitized area, which is then combined with the image from the video or picture to generate the final result image.

[0031] In a third aspect, the present invention provides an electronic device, comprising: a processor; and a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to perform the video and image desensitization method described in the first aspect above.

[0032] In a fourth aspect, the present invention provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to perform the video and image desensitization method according to the first aspect above.

[0033] The technical effects of this invention are as follows:

[0034] Currently, the desensitization methods used for vehicle videos and images are basically pixel erasure or color block replacement in sensitive areas. This invention, however, uses motion blur to desensitize sensitive areas (faces or license plates) while maintaining the overall layout and aesthetics of the image. Furthermore, to address the reversibility of motion blur, this invention segments the sensitive areas and uses different motion blur algorithms in different directions for desensitization of the sensitive areas. Thus, while desensitizing the sensitive areas (faces or license plates), the desensitized image is irreversible due to the segmentation of the sensitive areas and the motion blur processing in different vector directions. This achieves the dual effect of image desensitization and maintaining the image layout and aesthetics. Attached Figure Description

[0035] The accompanying drawings used in the following description of the embodiments or prior art will be briefly introduced. 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.

[0036] Figure 1This is a flowchart of the currently used video and image desensitization methods;

[0037] Figure 2 This is a network structure diagram of the currently used video and image desensitization methods;

[0038] Figure 3 This is an example image of sensitive area desensitization (color block erasure) using currently employed video and image desensitization methods.

[0039] Figure 4 This is a schematic diagram illustrating an embodiment of the overall process of the irreversible motion-blurred video and image desensitization method proposed in this invention;

[0040] Figure 5 This is a schematic diagram of an embodiment of the image recognition and desensitization process of the irreversible motion-blurred video and image desensitization method proposed in this invention;

[0041] Figure 6 This is a schematic diagram of an embodiment of the slice segmentation process of the irreversible motion-blurred video and image desensitization method proposed in this invention;

[0042] Figure 7 This is an example image of a face (sensitive area, or license plate, etc.) sliced ​​using the irreversible motion-blurred video and image desensitization method proposed in this invention;

[0043] Figure 8 This is a flowchart illustrating the motion blurring process of the irreversible motion-blurred video and image desensitization method proposed in this invention.

[0044] Figure 9 This is an example diagram of the irreversible motion-blurred video and image desensitization device proposed in this invention;

[0045] Figure 10 This is a schematic diagram of an electronic device. Detailed Implementation

[0046] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. It should be understood that the present invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. These embodiments are provided to provide a more thorough and complete understanding of the present invention. The accompanying drawings and embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Based on the embodiments of the present invention, technical solutions obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0047] It should be noted that the illustrations provided in the following embodiments are merely schematic representations of the basic concept of the present invention. Therefore, the drawings only show components relevant to the present invention and are not drawn according to the actual number, shape, and size of components in implementation. In actual implementation, the form, quantity, and proportion of each component can be arbitrarily changed, and the component layout may be more complex. Furthermore, it should be understood that the steps described in the method embodiments of the present invention can be performed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0048] This invention primarily relates to sensitive area occlusion during the desensitization process of sensitive areas (such as faces and license plates) in video and image transmission, particularly concerning the compliance processing of videos and images captured by vehicles. To protect the identity and vehicle information of external vehicles and personnel captured during vehicle photography, and to ensure the overall appearance and layout of the image, it is necessary to blur facial and license plate information. Sensitive areas processed using a motion blur algorithm make faces appear as if they were being photographed in motion, thus desensitizing sensitive areas such as faces and license plates. Its main difference from color block occlusion or pixel erasure is that it maintains the overall visual appeal of the image while desensitizing.

[0049] When taking photos of a vehicle remotely, the camera captures sensitive information such as faces and license plate numbers of pedestrians and other vehicles outside the vehicle. To protect the privacy of individuals appearing in the images, sensitive areas in the images need to be masked before the photos can be viewed.

[0050] See Figure 1 The current process for de-identifying vehicle videos and images is as follows:

[0051] 1. Video or Image Input: Transmits video or images captured by the vehicle's body equipment to the vehicle-side data processing equipment. Data includes video files, image files, and image data.

[0052] 2. Video or image preprocessing: After the transmission of video and image data is completed, the data is transcoded, decoded, or frame extracted for subsequent sensitive area location and desensitization processing.

[0053] 3. Sensitive region localization: Using a deep convolutional network and a high-recall detector FPN, the fixed network structure parameters obtained through data training are used to achieve accurate localization of face / license plate regions in video frames.

[0054] The network structure is as follows: Figure 2As shown, it detects each frame of video image obtained in the preprocessing stage and outputs the spatial information of each face and license plate in each frame of the target video, specifically including frame number, face position information and license plate position information.

[0055] 4. Desensitization Processing: Face and license plate location information in each frame of the video is replaced using uniform color blocks. This color block replacement directly erases pixel-level data from the original image, ensuring that the erased data is irreversible and unrecoverable. A diagram illustrating face desensitization is shown below. Figure 3 As shown.

[0056] 5. Post-processing: After desensitizing each frame of the video, the video frames can be converted into video in an orderly manner according to the frame number, based on the original video encoding, frame rate information, etc. The format of the desensitized and converted video should be consistent with the original video format.

[0057] 6. Output: Transmit the anonymized video or images to the enterprise's remote information service platform.

[0058] above Figure 1 , Figure 2 and Figure 3 This is derived from "Requirements and Methods for Desensitization Technology of Video and Image Transmission in Automobiles".

[0059] As can be seen from the above processing methods and illustrations, using color block replacement and pixel erasure will destroy the overall appearance of the image. Therefore, in image desensitization processing, a better approach is to preserve the overall appearance of the image while completing the irreversible desensitization process. That is, the focus is on covering sensitive areas (faces or license plates, etc.) while preserving the image's appearance and ensuring that the desensitized image is irreversible.

[0060] Example 1:

[0061] To address the above issues, the following provides an embodiment of the irreversible motion-blurred video and image desensitization method of the present invention. The following implementation process focuses on the image desensitization processing stage after sensitive area identification.

[0062] This embodiment employs a motion blur algorithm to blur sensitive areas (faces or license plates), preserving the overall appearance of the image while achieving desensitization. The key points are preserving the image's appearance while obscuring sensitive areas (faces or license plates, etc.), and ensuring that the desensitization process is irreversible. (See [link to previous document]). Figure 4 The overall processing flow of this embodiment is as follows:

[0063] Step 1. Identify the sensitive areas:

[0064] First, locate the sensitive areas of the video or image. After locating the sensitive areas (such as faces or license plates), obtain the sensitive areas.

[0065] For locating sensitive areas, existing technologies offer numerous solutions. For instance, a deep convolutional network (DCNN) and a high-recall detector (FPN) can be used. By training data and establishing fixed network structure parameters, precise location of face / license plate regions within video frames can be achieved. Figure 2 As shown.

[0066] Step 2. Slicing and dividing:

[0067] Based on the detection of sensitive regions, the sensitive regions are segmented into slices, such as... Figure 6 In this process, the sensitive regions of the image are segmented into slices, and the slice segmentation method is not limited to... Figure 6 The method allows for random cutting.

[0068] Step 3. Motion blur:

[0069] Based on image segmentation, motion blur algorithms are applied to blur each segmented portion. For example... Figure 7 As shown, the direction of motion blur is not limited to the specified direction shown in the figure; in terms of the degree of blurring, a random blurring method can be used; random slicing, random direction, and random blurring degree can be used to blur the desensitized area.

[0070] Step 4. Image Combining:

[0071] The processed desensitized regions are merged to obtain a composite image of the desensitized regions. Then, the composite image of the desensitized regions is merged back into the original video image frame or the original image to generate the final desensitized result image.

[0072] The above processing procedure is specifically as follows: Figure 5 As shown, the first image is a photo taken of a vehicle. The second image shows the process of detecting sensitive areas (faces) in the photo. To identify the sensitive areas that need processing, various face or text recognition algorithms can be used to automatically detect faces and license plates, thus identifying faces in the photo. The third image shows the desensitization processing of sensitive areas (faces). Once sensitive areas, such as faces and license plates, are detected, desensitization methods can be applied to desensitize the person. In practical applications, pixel erasure and color block replacement are commonly used. In this embodiment, the motion blur algorithm demonstrated above is used to blur the face, and the processed image is then synthesized. The fourth image shows the output of the photo after the sensitive areas have been processed.

[0073] This processing method utilizes the principle of motion blur. Motion blur occurs when the subject shifts during the camera's exposure time, and algorithms related to motion blur effects apply this principle to blur images. Compared to other desensitization methods, motion blur can make sensitive areas such as faces or license plates appear more natural and aesthetically pleasing in the processed image. Due to the commonality of motion blur, researching the restoration of motion-blurred images is a frequent research direction, and several motion blur recovery algorithms, such as the Wiener filter, have been developed to restore the blurred effect. However, in this invention, when applied to image desensitization in image processing, it is necessary to consider that the desensitized image must be irreversible.

[0074] To address this issue, this invention segments the sensitive region and applies different motion blurring methods to each segmented part. Specifically, by changing the direction and displacement of the vector field in the sensitive region, the sensitive region is blurred in different directions with personalized degrees of blurring. This achieves desensitization while maintaining the overall layout and aesthetics of the image, and prevents the desensitized areas from being reverse-engineered. See [link to related document]. Figure 6 .

[0075] Specifically, in step 2, the sensitive region is sliced. The general requirement for slicing is that the sensitive region is sliced, and each slice is not cut in a fixed way or is cut randomly in a fixed way. For example, by adjusting the width of horizontal and vertical cuts and the radius of ring cuts, the random combination of desensitized images is ensured, so that the desensitized images cannot be restored by using Wiener filtering or similar methods to restore motion-blurred images.

[0076] In practice, based on the characteristics of the image, at least the following segmentation methods can be selected, for example:

[0077] 2.1 Slicing based on specific values: This method segments the pixels in the image according to specific values, such as grayscale value, mean, variance, etc., and is more suitable for high-contrast images.

[0078] 2.2 Region-based segmentation: This method divides the pixels in an image into several parts and performs segmentation based on these regions. This method is suitable for images with complex backgrounds and high levels of noise.

[0079] 2.3 Edge-based segmentation: By detecting contours in the image, edge regions of parts of the image are located. This method is less effective for images with complex backgrounds and is suitable for images with relatively obvious edges. The edge refers to the feature or contour of the image to be de-identified (e.g., a face, license plate).

[0080] Therefore, in practical applications, it is necessary to select the slice segmentation method based on the usage scenario and the characteristics of each method.

[0081] The above-mentioned region-based segmentation method is highly usable in complex environments and allows for the introduction of randomness during processing; therefore, region-based segmentation is recommended. Specific methods include horizontal and vertical width cutting, fan-shaped arc cutting around a central region, and ring-shaped cutting with a circle as the center radius. Combinations of these methods can also be made. Sensitive regions after segmentation need to be recombined after desensitization; therefore, the slicing scheme and the recombining scheme must be continuous. In this embodiment, as shown... Figure 7 As shown, the sensitive area was divided into four parts by horizontal and vertical slicing.

[0082] Furthermore, in step 3, the motion blur processing method applied to each part after slice segmentation is as follows: the general requirement for using the motion blur algorithm is to blur the sensitive areas. Typically, the line integral convolution (LIC) motion blur algorithm or the canonical mapping function motion blur algorithm can be used to process the segmented image. The following explains the use of these two algorithms respectively:

[0083] 3.1 Line Integral Convolution (LIC) Motion Blur Algorithm:

[0084] This algorithm effectively represents an image as a two-dimensional vector field, which reflects the velocity and direction of each point. Noise is added to the image based on this two-dimensional vector field, and white noise texture is bidirectionally and symmetrically convolved along the streamline direction using a one-dimensional low-pass convolution kernel to finally synthesize an appropriate amount of texture. Therefore, for this invention, based on the image slicing in step 2 above, random noise can be added to each part of the segmented image to achieve image blurring. For example, the Line Integral Convolution (LIC) motion blur algorithm can be used for each part, performing blurring operations in different directions, as shown below. Figure 8 As shown in the diagram, the arrows indicate the movement directions of each part, which are all different and can be randomly selected. Furthermore, since our fundamental requirement is the irreversibility of image desensitization, the direction of the computation vector can be randomly generated, such as 1-upward, 2-downward, 3-leftward, 4-rightward, and operations can be performed using directions like 1->3->4->2 or 4->2->1->3.

[0085] 3.2 Standardized Mapping Function Motion Fuzzy Algorithm:

[0086] For this algorithm, the mapping function is as follows: for any two distinct elements x1 and x2 in X, x1 ≠ x2, the image obtained through the mapping function y = f(x1) ≠ (x2). When processing the segmented sensitive region image, this algorithm adjusts the input value of the mapping function differently, thereby making the displacement of the processed image in the direction of motion random, thus achieving the effect that the entire sensitive image cannot be restored with the same parameters.

[0087] Therefore, based on the cutting method in step 2 above and the characteristics of the two blurring algorithms, one or a combination of these two algorithms can be used when performing image desensitization processing. Using the normalized mapping function motion blur algorithm for image processing, and incorporating randomization of the function's independent variables and improving processing efficiency, can improve the randomness and efficiency of desensitization.

[0088] In practical applications, the motion blur processing scheme can be adapted to different parameters for each segment. Alternatively, different algorithms with different parameters can be used for different slices. The applied blur algorithm is not limited to a specific motion blur algorithm.

[0089] Therefore, considering the above slicing and blurring processing schemes, as well as performance resource consumption and efficiency, we can use the four directions mentioned in the slicing scheme of this embodiment to randomly generate the order of the four directions, and then use the line integral convolution (LIC) motion blurring algorithm on each part. Considering performance consumption, we can use a single algorithm for blurring processing.

[0090] As can be seen from the above embodiments, the present invention employs a motion blur algorithm to apply different directions and degrees of blurring to different slices based on slicing the desensitized areas of the image. This maintains the overall layout and aesthetics of the graphic while ensuring security, making the desensitized image unrecoverable.

[0091] Finally, for step 4, the processed desensitized regions are merged, and a composite image of the desensitized regions is output. The desensitized image is then combined with the original image and output to generate the final result image. This step can be performed using various existing methods.

[0092] The general requirement for image merging is that the merged images after segmentation should have completeness.

[0093] Common methods of merging include regional growth, regional merging, regional segmentation and merging, and statistical regional merging.

[0094] Because this invention incorporates randomness in the image segmentation and motion blur processes, it employs a reverse merging method to ensure the integrity of the image.

[0095] Example 2:

[0096] In another embodiment, an apparatus is provided to implement the above-described irreversible motion-blurred video and image desensitization method. To implement the above method, the desensitization apparatus includes the following functional modules, such as... Figure 9 As shown:

[0097] The sensitive area acquisition module is used to acquire sensitive areas of videos or images. Specifically, it first locates sensitive areas (such as faces or license plates) in the video or image, and then acquires these sensitive areas as targets for subsequent de-identification processing.

[0098] The slice segmentation module is used to segment sensitive regions of an image. Specifically, based on the detection of sensitive regions, it selects an appropriate slice segmentation method according to the image's features, such as slice segmentation based on specific values, segmentation based on regions as features, or segmentation based on edges as dimensions.

[0099] The motion blur processing module is used to blur each part of the slice using different motion blur methods. Specifically, it can use the line integral convolution (LIC) motion blur algorithm or the canonical mapping function motion blur algorithm to process the sliced ​​image.

[0100] The image compositing module is used to merge the processed parts to obtain a composite image of the desensitized area, which is then combined with the image from the video or picture to generate the final result image.

[0101] Example 3:

[0102] In a further embodiment, an electronic device is proposed, comprising: a processor; and a memory storing a program, wherein the program includes instructions that, when executed by the processor, cause the processor to perform the aforementioned irreversible motion-blurred video and image desensitization method. Figure 10 As shown, the electronic device 100 includes a processor 101 and a memory 103. The processor 101 and the memory 103 are connected, for example, via a bus 102.

[0103] The structure of the electronic device 100 does not constitute a limitation on the embodiments of the present invention.

[0104] Processor 101 may be a CPU, a general-purpose processor, a DSP, an ASIC, an FPGA, or other programmable logic device, transistor logic device, hardware component, or any combination thereof. It may implement or execute the various exemplary logic blocks, modules, and circuits described in connection with this disclosure. Processor 101 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0105] Bus 102 may include a pathway for transmitting information between the aforementioned components. Bus 102 may be a PCI bus or an EISA bus, etc. Bus 102 may be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used to represent it in the figure, but this does not mean that there is only one bus or one type of bus.

[0106] The memory 103 may be a ROM or other type of static storage device capable of storing static information and instructions, RAM or other type of dynamic storage device capable of storing information and instructions, or it may be an EEPROM, CD-ROM or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage medium or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0107] The applicant of this invention has provided a detailed description of the embodiments of the invention in conjunction with the accompanying drawings. However, those skilled in the art should understand that the above embodiments are merely preferred embodiments of the invention. The detailed description is only to help better understand the spirit of the invention and is not intended to limit the scope of protection of the invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, but without departing from the spirit of the invention or exceeding the scope defined by the appended claims.

[0108] Based on the embodiments of the present invention described above, and through the above description, those skilled in the art can make various changes and modifications without departing from the technical concept of the present invention. The scope of the present invention is not limited to the specific embodiments described, but should be determined according to the scope of the claims.

Claims

1. A method for desensitizing videos and images, characterized in that, include: To capture the sensitive areas of a video or image; Slice the sensitive regions of the image into segments; For each part segmented by the slice, motion blurring is applied to blur it in different directions and to different degrees. The processed parts are merged to obtain a composite image of the desensitized area, which is then merged back into the original video image frame or the original image to generate the final desensitized result image. The process of performing motion blurring on each part of the slice segment in different directions and to different degrees is as follows: by changing the direction and displacement of the vector field of each part of the slice segment, the sensitive area is blurred in different directions with different degrees of blurring, wherein at least two parts have different directions and displacement of the vector field. When performing motion blurring, different directions are used for blurring operations on each part, and the operation vector direction is randomly generated. The selection of motion fuzzing algorithms for each part involves randomly selecting the same or different computational algorithms. The slicing includes cutting in random, non-fixed rules, different directions, different sizes and / or different shapes.

2. The video and image desensitization method according to claim 1, characterized in that, The motion blur algorithm can be either the line integral convolution LIC motion blur algorithm or the canonical mapping function motion blur algorithm.

3. The video and image desensitization method according to claim 1 or 2, characterized in that, The slice segmentation includes: Slicing based on specific values, where the specific values ​​are grayscale values, average values, or variances; Slicing based on regions involves dividing an image into several regions by pixels, and then dividing the image into several parts based on those regions. Edge-based slicing segmentation involves locating and segmenting edge regions of an image by detecting contours within the image.

4. The video and image desensitization method according to claim 3, characterized in that, The region-based slicing segmentation employs any one of the following methods or a combination thereof: horizontal and vertical cutting, rotational cutting around the central region, or circular cutting centered on a circle.

5. A video and image desensitization device, characterized in that, include: The sensitive area acquisition module is used to acquire the sensitive areas of a video or image. The slice segmentation module is used to slice and segment sensitive regions of an image; The motion blur processing module is used to perform motion blur processing on each part of the sliced ​​area in different directions and to different degrees. The image compositing module is used to merge the processed parts to obtain a composite image of the desensitized area, and then merge it back into the original video image frame or the original image to generate the final desensitized result image. The motion blurring processing module is configured to blur different parts of the sensitive region in different directions with different degrees of blurring by changing the direction and displacement of the vector fields of each part segmented by the slice, wherein at least two parts have different directions and displacement of the vector fields; The motion blurring processing module performs blurring operations in different directions during motion blurring processing, and the operation vector direction is randomly generated. The motion blurring processing module randomly selects the same or different motion blurring algorithms for each part. The slicing module is configured to cut in a random, non-fixed, different direction, different size and / or different shape manner.

6. The video and image desensitization device according to claim 5, characterized in that, The motion blur algorithm is selected from either the line integral convolution LIC motion blur algorithm or the canonical mapping function motion blur algorithm.

7. The video and image desensitization device according to claim 5 or 6, characterized in that, The slicing method used by the slicing module includes: Slicing based on specific values, where the specific values ​​are grayscale values, average values, or variances; Slicing based on regions involves dividing an image into several regions by pixels, and then dividing the image into several parts based on those regions. Edge-based slicing segmentation involves locating and segmenting edge regions of an image by detecting contours within the image.

8. The video and image desensitization device according to claim 7, characterized in that, The region-based slicing segmentation employs any one of the following methods or a combination thereof: horizontal and vertical cutting, rotational cutting around the central region, or circular cutting centered on a circle.

9. An electronic device, comprising: processor; And a memory for storing a program, characterized in that the program includes instructions that, when executed by the processor, cause the processor to perform the video and image desensitization method according to any one of claims 1-4.

10. A non-transitory computer-readable storage medium storing computer instructions, characterized in that, The computer instructions are used to cause the computer to execute the video and image desensitization method according to any one of claims 1-4.

Images