Image compression device and image compression method

Abstract

This image compression device comprises: an extraction unit that extracts a region being protected, which includes a protection subject in personal information, from an image; and a compression unit that compresses the region being protected at a compression ratio higher than that for regions in the image other than the region being protected, thereby compressing the image.

Description

[Technical Field]

[0001] This disclosure relates to an image compression device and an image compression method. This application claims priority to Japanese Application No. 2022-171929 filed on October 27, 2022, and incorporates by reference all of the contents of said Japanese application. [Background technology]

[0002] BACKGROUND ART Utilization of images captured by various cameras, such as in-vehicle cameras mounted on automobiles, cameras mounted on mobile robots, and surveillance cameras installed in factories or on roads, is becoming more widespread.

[0003] For example, these images are expected to be used for map making, road maintenance, area marketing, or production line monitoring.

[0004] On the other hand, there is a possibility that personal information such as people's faces, house nameplates, the interior of houses, or car license plates may be captured in the video. Each country has its own rules regarding the handling of such personal information. For example, in the European Union (EU), such information is protected under the General Data Protection Regulation (GDPR).

[0005] Artificial intelligence (AI) is used to remove the protected object from the video (see, for example, Non-Patent Document 1 and Non-Patent Document 2).

[0006] Conventional AI detects areas of video that contain faces or license plates where personal information needs to be protected, and removes personal information from the video by mosaic-processing the detected areas or by replacing the faces or license plates with faces or license plates generated by AI. [Prior art documents] [Non-patent literature]

[0007] [Non-Patent Document 1] “Masking-AI”, [online], 2021, TRUST SMITH Co., Ltd., [Retrieved August 8, 2022], Internet<URL:https: / / www.trustsmith.net / masking-ai / > [Non-patent document 2] “brighter AI”, [online], Macnica Holdings Inc., [Retrieved August 8, 2022], Internet<URL:https: / / www.macnica.co.jp / business / dx / manufacturers / brighterai / > Summary of the Invention

[0008] An image compression device according to one aspect of the present disclosure includes an extraction unit that extracts a protected area from an image, which is an area that includes personal information to be protected, and a compression unit that compresses the image by compressing the protected area at a higher compression rate than areas other than the protected area in the image.

[0009] The present disclosure may be realized as an image compression method including, as steps, characteristic processes executed by an image compression device. It may also be realized as a computer program for causing a computer to function as a characteristic processing unit included in an image compression device, or as a computer program for causing a computer to execute characteristic steps included in an image compression method. It goes without saying that such a computer program may be distributed via a computer-readable, non-transitory recording medium such as a CD-ROM (Compact Disc-Read Only Memory) or a communication network such as the Internet. The present disclosure may also be realized as a semiconductor integrated circuit that implements part or all of the image compression device, or as an image processing system including an image compression device. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a diagram illustrating an example of the overall configuration of an image processing system according to the first embodiment. [Figure 2] FIG. 2 is a diagram showing an example of an image output by the camera. [Figure 3] FIG. 3 is a block diagram showing an example of the configuration of the image compression device according to the first embodiment. [Figure 4] FIG. 4 is a diagram in which the image shown in FIG. 2 is divided into a plurality of blocks. [Figure 5] FIG. 5 is a diagram showing an example of a region to be protected in an image. [Figure 6] FIG. 6 is a diagram showing an example of the result of extraction of the protection target area by the protection target area extraction unit. [Figure 7] FIG. 7 is a block diagram illustrating an example of the configuration of the server according to the first embodiment. [Figure 8] FIG. 8 is a flowchart showing an example of a procedure of processing executed by the camera and the image compression device. [Figure 9] FIG. 9 is a flowchart illustrating an example of a procedure of a process executed by the server. [Figure 10] FIG. 10 is a block diagram illustrating an example of the configuration of an image compression device according to the second embodiment. [Figure 11] FIG. 11 is a flowchart showing an example of a procedure of processing executed by the camera and the image compression device. [Figure 12] FIG. 12 is a diagram showing an example of the result of extraction of the protection target area and the protection target size by the protection target area extraction unit. [Figure 13] FIG. 13 is a flowchart illustrating an example of a procedure of processing executed by the camera and the image compression device. [Figure 14] FIG. 14 is a flowchart illustrating an example of a procedure of processing executed by the camera and the image compression device. [Figure 15] FIG. 15 is a block diagram illustrating an example of the configuration of a server according to the fifth embodiment. [Figure 16] FIG. 16 is a flowchart illustrating an example of a procedure of a process executed by the server. DETAILED DESCRIPTION OF THE INVENTION

[0011] [Problem to be solved by this disclosure] Conventional AI can remove personal information, but it does not compress the video, which means the video data volume is large.

[0012] The present disclosure has been made in consideration of the above-mentioned problems, and aims to provide an image compression device and an image compression method that can remove personal information while reducing the amount of image data.

[0013] [Effects of this disclosure] According to the present disclosure, it is possible to remove personal information while reducing the amount of image data.

[0014] [Summary of the embodiments of the present disclosure] First, an outline of the embodiments of the present disclosure will be listed and described. (1) An image compression device according to one embodiment of the present disclosure includes an extraction unit that extracts a protected area from an image, the protected area being an area that includes personal information to be protected, and a compression unit that compresses the image by compressing the protected area at a higher compression rate than areas other than the protected area in the image.

[0015] For example, the compression unit compresses the protection target area at a high compression rate that makes it impossible to recognize personal information such as a face or license plate even if the compressed protection target area is expanded. This allows image compression and removal of personal information to be performed simultaneously. In other words, personal information can be removed while reducing the amount of image data.

[0016] (2) In the above (1), the extraction unit may extract the area to be protected by determining whether each of a plurality of blocks obtained by dividing the image corresponds to the area to be protected.

[0017] By determining whether a block corresponds to a protection target area, the process of determining the protection target area can be performed efficiently. At the same time, since the image can be compressed on a block-by-block basis, image compression and removal of personal information can be performed efficiently.

[0018] (3) In the above (1) or (2), the image compression device may further include a masking processing unit that performs a masking process to remove personal information from the protected area, and the compression unit may compress the image after the masking process.

[0019] For example, if the size of the protection target, such as a face or license plate included in the protection target area, is large, personal information may be read from the decompressed image. However, by performing a masking process, the personal information can be reliably removed and the amount of image data can be reduced. Note that the size of the protection target does not refer to the actual size of the protection target (actual size in three-dimensional space), but rather to its size in the image. For example, the size of the protection target refers to the number of pixels that the protection target area occupies in the image.

[0020] (4) In the above (3), the masking processing unit may perform the masking process on the protection target area based on the size of the protection target included in the protection target area.

[0021] For example, if the size of the object to be protected, such as a face or license plate, included in the protection area is larger than a certain value, masking processing can be performed, and in other cases masking processing cannot be performed. This makes it possible to prevent unnecessary masking processing, reduce the amount of image data, and remove personal information.

[0022] (5) In any of (1) to (4) above, the compression unit may determine a compression rate for compressing the protection target area based on the size of the protection target included in the protection target area.

[0023] For example, the larger the size of the protection target, such as a face or license plate included in the protection target area, the higher the compression rate can be, which makes it possible to reduce the amount of image data while reliably removing personal information.

[0024] (6) In (4) or (5) above, the size of the protection object may be determined based on the area of ​​the protection object included in the protection object region and the area of ​​each block obtained by dividing the image.

[0025] For example, if the area of ​​the object to be protected is larger than the area of ​​the block, masking can be performed, and if not, masking can be omitted. This prevents unnecessary masking, reduces the amount of image data, and removes personal information.

[0026] (7) In any one of (1) to (6) above, the image compression device may further include an output unit that outputs position information of the protection target area in the image.

[0027] For example, a server that acquires the compressed image and the location information of the protected area can decompress the image while changing the compression rate based on the location information. Furthermore, if no masking process has been performed, the server can mask the protected area based on the location information. This allows the server to reliably remove personal information before performing other image processing.

[0028] (8) An image compression method according to another embodiment of the present disclosure includes extracting a protected area from an image, the protected area being an area containing personal information to be protected, and compressing the image by compressing the protected area at a higher compression rate than areas other than the protected area in the image.

[0029] This configuration includes the characteristic processing steps of the above-described image compression device, and therefore provides the same functions and effects as the above-described image compression device.

[0030] [Details of the embodiments of the present disclosure] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings. Note that each of the embodiments described below represents a specific example of the present disclosure. The numerical values, shapes, materials, components, component placement and connection configurations, steps, and step order shown in the following embodiments are examples and do not limit the present disclosure. Furthermore, among the components in the following embodiments, components not recited in independent claims are components that can be added arbitrarily. Furthermore, each figure is a schematic diagram and is not necessarily a precise illustration.

[0031] The same components are denoted by the same reference numerals, and their functions and names are also the same, so their explanations will be omitted where appropriate.

[0032] <Embodiment 1> [Overall configuration of image processing system] FIG. 1 is a diagram illustrating an example of the overall configuration of an image processing system according to the first embodiment. The image processing system 100 includes an image compression device 1 and a server 2 connected to each other via a network 3 such as the Internet, a camera 1A connected to the image compression device 1, and a display device 2A connected to the server 2. The image compression device 1 and the server 2 are connected to the network 3 by wire or wirelessly.

[0033] Camera 1A is, for example, an in-vehicle camera mounted on an automobile. However, camera 1A is not limited to an in-vehicle camera. For example, camera 1A may be a camera mounted on a mobile robot, a surveillance camera attached to factory equipment, or an infrastructure camera installed on the roadside to monitor traffic flow.

[0034] Camera 1A captures an image of a subject and outputs the image of the subject. The image may be a time-series image, i.e., a video. Camera 1A may be provided inside image compression device 1 and may be included as a component of image compression device 1.

[0035] 2 is a diagram showing an example of an image output by the camera 1A. The image shown in FIG. 2 includes a person 41, a person 42, and a person 43. , Person 44 and Person 45 is shown. The image compression device 1 compresses an image captured by the camera 1A and transmits the compressed image to the server 2.

[0036] The server 2 receives a compressed image from the image compression device 1 via the network 3 and decompresses the received compressed image. The server 2 performs predetermined image processing on the decompressed image. For example, the server 2 detects the position of an object included in the image, the circumscribing area of ​​the object, etc. by image processing.

[0037] The display device 2A is configured with a liquid crystal display or an organic EL (electroluminescence) display, and displays the image after decompression by the server 2, the result of image processing by the server 2, or the like.

[0038] [Configuration of image compression device 1] FIG. 3 is a block diagram showing an example of the configuration of the image compression device 1 according to the first embodiment. The image compression device 1 is connected to a camera 1A, and includes a protection target area extraction unit 11, an area-specific compression unit 12, an image transmission unit 13, and a storage device .

[0039] The image compression device 1 may be configured as a computer system including a processor such as a CPU (Central Processing Unit) and a storage device 14 such as a volatile memory or a non-volatile memory. Each of the processing units 11, 12, and 13 performs its function by executing a computer program stored in the storage device 14 on the processor.

[0040] Protection target area extraction unit 11 receives from camera 1 A an image captured by camera 1 A. Protection target area extraction unit 11 extracts a protection target area, which is an area including personal information to be protected, from the received image.

[0041] For example, if the unit of an image to be judged as to whether it contains a protection target is a block, the protection target area extraction unit 11 divides the image into multiple blocks and judges for each block whether it contains a predetermined protection target. As a result, the area-by-area compression unit 12 (described later) extracts the protection target area to be compressed at a higher compression rate than areas other than the protection target area in the image.

[0042] The processing of the protection target area extraction unit 11 will be described in detail below. Figure 4 is a diagram showing the image shown in Figure 2 divided into a plurality of blocks. Here, an example is shown in which one image is divided into 360 blocks of 15 rows and 24 columns. As shown in Figure 4, block numbers from 1 to 360 are assigned to each block in raster scan order from the top left to the bottom right of the image. However, the size and number of blocks are not limited to those shown in Figure 4. The protection target to be determined by the protection target area extraction unit 11 is assumed to be, for example, a human face. However, the protection target is not limited to a human face and may be a license plate of an automobile or the like.

[0043] The protection target area extraction unit 11 uses a learning model to determine whether or not a human face is included in each block from block number 1 to block number 360. The learning model is, for example, a convolution neural network (CNN), a recurrent neural network (RNN), an autoencoder, etc. Using images of blocks containing various human faces (hereinafter referred to as "block images") as training data, learning of human faces progresses by a machine learning method such as deep learning, and parameters of the learning model are determined.

[0044] Specifically, the protection target area extraction unit 11 inputs an image consisting of 360 block images into a learning model. The learning model determines whether or not each block contains a protection target and outputs the determination result. The protection target area extraction unit 11 obtains the determination result from the learning model.

[0045] The protection target area extraction unit 11 determines a block including a protection target as a protection target area to be compressed at a higher compression rate than other blocks.

[0046] FIG. 5 is a diagram showing an example of a protection target area in an image. Protection target area extraction unit 11 extracts protection target areas from the image shown in FIG. 2 by classifying blocks containing protection targets that are adjacent to each other within four or eight neighborhoods as the same protection target area. As a result, protection target areas 41A, 43A, 44A, and 45A are extracted. Protection target area 41A is composed of a thick-framed block containing the face of person 41 and a thick-framed block containing the face of person 42. Protection target area 43A, protection target area 44A, and protection target area 45A are composed of thick-framed blocks containing the faces of person 43, person 44, and person 45, respectively.

[0047] Protection target area 41A is an area including blocks 80, 81, 102 to 105, 126, and 127. Protection target area 43A is an area including blocks 155, 156, 179, and 180. Protection target area 44A is an area including block 158. Protection target area 45A is an area including blocks 113, 114, 137, 138, 161, and 162.

[0048] FIG. 6 is a diagram showing an example of the result of extraction of the protection target area by the protection target area extraction unit 11. As shown in FIG. The table in Figure 6 indicates whether or not a protection target is included for each block number. That is, a block that includes a protection target is indicated by "1," and a block that does not include a protection target is indicated by "0." For example, the first block indicates that it does not include a protection target, while the 80th block indicates that it does include a protection target.

[0049] Referring back to FIG. 3, the protection target area extraction unit 11 outputs the extraction result of the protection target area to the area-by-area compression unit 12.

[0050] The area-specific compression unit 12 receives from the camera 1A the image captured by the camera 1A and compresses the image. More specifically, the area-specific compression unit 12 compresses the protection target area extracted by the protection target area extraction unit 11 at a higher compression rate than the area other than the protection target area in the image, thereby compressing the image.

[0051] Compression methods include, for example, H.264 / MPEG-4 AVC, or H.265 / MPEG-H HEVC, etc. The higher the compression rate, the lower the image quality of the expanded image (area).

[0052] In the above example, the area-specific compression unit 12 compresses the block images included in the protection target areas 41A, 43A, 44A, and 45A at a relatively higher compression rate than the other block images (for example, the block image of the first block), and compresses the other block images at a relatively lower compression rate. Here, the compression rate for compressing the block images included in the protection target areas is preferably set in advance to a compression rate that makes it impossible to recognize personal information such as a face or license plate even if the protection target area is expanded.

[0053] The area-by-area compression unit 12 adds position information of the area to be protected within the image to the compressed image. An example of this position information is the extraction result shown in FIG.

[0054] The area-specific compression unit 12 writes the compressed image with the position information of the protection target area added to the storage device 14. The area-specific compression unit 12 also outputs the compressed image with the position information of the protection target area added to the image transmission unit 13.

[0055] The image transmitting unit 13 transmits the compressed image to which the position information of the protection target area acquired from the area-by-area compressing unit 12 has been added via the network 3 to the server 2 .

[0056] [Server 2 configuration] FIG. 7 is a block diagram showing an example of the configuration of the server 2 according to the first embodiment. The server 2 includes an image receiving unit 21, an image decompression unit 22, an image processing unit 23, a display control unit 24, and a storage device 25.

[0057] The server 2 may be configured as a computer system including a processor such as a CPU and a storage device 25 such as a volatile memory or a non-volatile memory. Each of the processing units 21, 22, 23, and 24 performs its function by executing a computer program stored in the storage device 25 on the processor.

[0058] The image receiving unit 21 receives the compressed image to which the position information of the area to be protected has been added from the image compressing device 1 via the network 3 , and outputs the image to the image decompressing unit 22 .

[0059] The image decompression unit 22 decompresses the compressed image received by the image receiving unit 21 to restore the original image, and outputs the restored image to the image processing unit 23. Specifically, the image decompression unit 22 references the position information of the protection target area and decompresses the block images included in the protection target area using a decompression method corresponding to the compression method used for the block images included in the protection target area. The image decompression unit 22 also references the position information and decompresses the block images not included in the protection target area using a decompression method corresponding to the compression method used for the block images not included in the protection target area. In the example shown in FIG. 5, even if the block images included in the protection target areas 41A, 43A, 44A, and 45A are restored by decompression, a person or computer cannot recognize personal information such as faces.

[0060] The image processing unit 23 performs predetermined image processing on the decompressed image. For example, the image processing unit 23 detects a circumscribed rectangular area of ​​an object included in the image. The image processing unit 23 may also perform area segmentation processing to extract the area of ​​the object included in the image in pixel units. Furthermore, the image processing unit 23 may also perform skeleton estimation processing to estimate the skeleton of the object included in the image.

[0061] The image processing unit 23 outputs the result of the image processing to the display control unit 24. The image processing unit 23 also writes the result of the image processing into the storage device 25. The display control unit 24 displays the image processing results obtained from the image processing unit 23 on the display device 2A.

[0062] [Processing Procedure of Image Compression Device 1] FIG. 8 is a flowchart showing an example of the procedure of the process executed by the camera 1A and the image compression device 1. The camera 1A generates an image by capturing an image of a subject (step S101), thereby obtaining an image such as that shown in FIG.

[0063] The protection target area extraction unit 11 extracts protection target areas, which are areas including personal information to be protected, from the image captured by the camera 1A (step S102). For example, protection target areas 41A, 43A, 44A, and 45A as shown in FIG. 4 are extracted.

[0064] The area-specific compression unit 12 sets a higher compression rate for the protection target area extracted by the protection target area extraction unit 11 than for the area other than the protection target area in the image, and compresses the image at the set compression rate (step S103).

[0065] The area-by-area compression unit 12 adds the position information of the protection target area extracted in the process of step S102 to the compressed image generated in the process of step S103, and writes the image to the storage device 14 (step S104).

[0066] The image transmitting unit 13 obtains the compressed image to which the position information of the protection target area has been added from the area-by-area compressing unit 12, and transmits the image to the server 2 (step S105).

[0067] [Server 2 processing procedure] FIG. 9 is a flowchart illustrating an example of a procedure of a process executed by the server 2. The image receiving unit 21 receives the compressed image to which the position information of the area to be protected has been added from the image compressing device 1 via the network 3 (step S201).

[0068] The image decompression unit 22 decompresses the compressed image received by the image receiving unit 21 to restore the original image (step S202).

[0069] The image processing unit 23 performs predetermined image processing on the expanded image (step S203).

[0070] The image processing unit 23 writes the result of the image processing into the storage device 25 (step S204).

[0071] The display control unit 24 acquires the result of the image processing from the image processing unit 23 and causes the display device 2A to display it (step S205).

[0072] According to the first embodiment, the image compression device 1 compresses the protection target area at a high compression rate that makes it impossible to recognize personal information such as a face or license plate even if the compressed protection target area is expanded. This allows image compression and removal of personal information to be performed simultaneously. In other words, personal information can be removed while reducing the amount of image data.

[0073] Furthermore, the image compression device 1 can efficiently perform the process of determining whether an image falls within a protection target area by determining whether each block corresponds to a protection target area. At the same time, the image compression device 1 can efficiently perform image compression and removal of personal information because it can compress an image in block units.

[0074] <Embodiment 2> In the second embodiment, an example in which a masking process is further performed on the protection target area will be described. The configurations of the image processing system 100 and the server 2 are the same as those in the first embodiment.

[0075] FIG. 10 is a block diagram showing an example of the configuration of an image compression device 1 according to the second embodiment. The image compression device 1 is connected to a camera 1A, and includes a protection target area extraction unit 11, an area-specific compression unit 12, an image transmission unit 13, a storage device 14, and a masking processing unit 15.

[0076] The image compression device 1 may be configured as a computer system including a processor such as a CPU and a storage device 14 such as a volatile memory or a non-volatile memory. Each of the processing units 11, 12, 13, and 15 performs its function by executing a computer program stored in the storage device 14 on the processor.

[0077] The configuration of the protection target area extraction unit 11 is the same as that of the first embodiment, except that the protection target area extraction unit 11 outputs the extraction result of the protection target area to the masking processing unit 15 in addition to the area-by-area compression unit 12.

[0078] The masking processor 15 receives from the camera 1A the image captured by the camera 1A. Based on the result of extracting the protection target area, the masking processor 15 performs a masking process to remove personal information from all protection target areas in the image captured by the camera 1A. The masking process includes a mosaic process to mosaic the image, a blurring process to blur the image, or a smoothing process to smooth the image. For example, the masking processor 15 performs the masking process by performing at least one of a mosaic process, a blurring process, and a smoothing process on the block image included in the protection target area. The masking processing unit 15 outputs the image after the masking processing to the region-by-region compression unit 12 .

[0079] The regional compression unit 12 compresses the image after the masking process. The regional compression unit 12 is the same as the regional compression unit 12 shown in the first embodiment, except that the image to be compressed is the image after the masking process. The configurations of the image transmission unit 13 and the storage device 14 are the same as those in the first embodiment.

[0080] FIG. 11 is a flowchart showing an example of the procedure of the process executed by the camera 1A and the image compression device 1. Camera 1A executes the process of step S101, and image compression device 1 executes the process of step S102. The processes of steps S101 and S102 are the same as the processes of steps S101 and S102 described with reference to FIG.

[0081] Based on the result of extraction of the protection target area, the masking processing unit 15 performs masking processing to remove personal information from the protection target area in the image captured by the camera 1A (step S111).

[0082] The area-specific compression unit 12 compresses the image after the masking process by compressing the protection target area extracted by the protection target area extraction unit 11 at a higher compression rate than the area other than the protection target area in the image after the masking process (step S112).

[0083] Thereafter, the image compression device 1 executes the processes of steps S104 and S105. The processes of steps S104 and S105 are the same as the processes of steps S104 and S105 described with reference to FIG.

[0084] For example, if the size of the image to be protected, such as a face or license plate included in the protected area, is large, there is a possibility that personal information may be read from the expanded image. However, by performing a masking process on the protected area, the personal information can be reliably removed and the amount of image data can be reduced.

[0085] <Embodiment 3> In the second embodiment, the entire area to be protected is always masked. However, it may be determined whether or not to mask the entire area to be protected based on the size of the area to be protected. In this disclosure, the size of the area to be protected does not refer to the actual size of the area to be protected (actual size in three-dimensional space), but rather to the size of the area in the image. For example, the size of the area to be protected refers to the number of pixels that the area to be protected occupies in the image. The configurations of the image processing system 100 and the server 2 are the same as those in the first embodiment.

[0086] The configuration of the image compression device 1 is the same as that of the second embodiment. However, as will be explained below, some of the configuration is different from that of the second embodiment.

[0087] Referring to FIG. 10, protection target area extraction unit 11 extracts a protection target area and the size of the protection target included in the protection target area from an image captured by camera 1A.

[0088] For example, the protection target area extraction unit 11 divides an image into multiple blocks, and for each block, determines whether it contains a predetermined protection target, and if it does, the size of the protection target within the image.

[0089] The processing of the protection target area extraction unit 11 will be described in detail below. The protection target to be determined by the protection target area extraction unit 11 is assumed to be, for example, a human face. However, the protection target is not limited to a human face and may be a license plate of an automobile or the like.

[0090] Referring to FIG. 4, the protection target area extraction unit 11 uses a learning model to determine whether or not a human face is included in each block from block number 1 to block number 360, and if a human face is included, the size of the human face. The learning model is, for example, a CNN, an RNN, or an AutoEncoder. Using block images containing various human faces as training data, learning of human faces and their sizes in images is carried out by a machine learning method such as deep learning, and parameters of the learning model are determined. The size of the protection target (the size of the human face) is determined based on the area of ​​the protection target in the image and the area of ​​the block. In particular, if the area of ​​the protection target is larger than the area of ​​the block, the size of the protection target is determined to be large, and if the area of ​​the protection target is equal to or smaller than the area of ​​the block, the size of the protection target is determined to be small.

[0091] Specifically, the protection target area extraction unit 11 inputs an image consisting of 360 block images into a learning model. The learning model determines whether or not a protection target is included in each block, and if a protection target is included, the size of the protection target, and outputs the determination result. The protection target area extraction unit 11 obtains the determination result from the learning model.

[0092] 5, protection target area extraction unit 11 extracts protection target areas 41A, 43A, 44A, and 45A from the image shown in Fig. 2. Furthermore, protection target area extraction unit 11 determines that the size of the protection target included in protection target areas 41A, 43A, and 45A is large, and determines that the size of the protection target included in protection target area 44A is small.

[0093] FIG. 12 is a diagram showing an example of the result of extraction of the protection target area and protection target size by the protection target area extraction unit 11. In FIG. The table shown in Figure 12 indicates, for each block number, whether or not a protection object is included, and if a protection object is included, the size of the protection object. That is, a block that includes a protection object is indicated by "1," and a block that does not include a protection object is indicated by "0." For example, the first block indicates that it does not include a protection object. The 80th block (a block included in the protection area 41A) indicates that it includes a protection object, and the size of the protection object is large. The 158th block (a block included in the protection area 44A) indicates that it includes a protection object, and the size of the protection object is small.

[0094] Referring again to FIG. 10, protection target area extraction unit 11 outputs the extraction results of the protection target area and the protection target size to area-by-area compression unit 12 and masking processing unit 15.

[0095] The masking processing unit 15 refers to the extraction results of the protection target area and the protection target size, and determines whether or not there is a protection target area including a large-sized protection target in the image captured by the camera 1A. For example, according to the extraction results of the protection target area and the protection target size shown in FIG. 12, the protection target flag of block number 80 is "1" and the protection target size is "large". Therefore, there is a protection target area including a large-sized protection target in the image.

[0096] If there is even one protection target area that includes a large-sized protection target, the masking processing unit 15 executes masking processing to remove personal information from all protection target areas in the image captured by the camera 1A.

[0097] The region-by-region compression unit 12 compresses the image after the masking process, as in the second embodiment. The region-by-region compression unit 12 adds position information of the region to be protected in the image to the compressed image. An example of this position information is the extraction result shown in FIG. 12.

[0098] The area-specific compression unit 12 writes the compressed image with the position information of the protection target area added to the storage device 14. The area-specific compression unit 12 also outputs the compressed image with the position information of the protection target area added to the image transmission unit 13.

[0099] The image transmitting unit 13 transmits the compressed image to which the position information of the protection target area acquired from the area-by-area compressing unit 12 has been added via the network 3 to the server 2 .

[0100] FIG. 13 is a flowchart showing an example of the procedure of the process executed by the camera 1A and the image compression device 1. Camera 1A executes the process of step S101, and image compression device 1 executes the process of step S102. The processes of steps S101 and S102 are the same as the processes of steps S101 and S102 described with reference to FIG.

[0101] Based on the result of extraction of the protection target area, the masking processing unit 15 determines whether or not the image captured by the camera 1A includes a protection target area that includes a large-sized protection target (step S121).

[0102] If the protection target area includes a large-sized protection target (YES in step S121), the masking processing unit 15 performs masking processing to remove personal information from all protection target areas in the image captured by camera 1A based on the extraction results of the protection target area and protection target size (step S122).

[0103] If a protection target area including a large-sized protection target is not included (NO in step S121), the process of step S122 is not executed.

[0104] Thereafter, the image compression device 1 executes the processes of steps S112, S104, and S105. The process of step S112 is the same as the process of step S112 described with reference to Fig. 11. The processes of steps S104 and S105 are the same as the processes of steps S104 and S105 described with reference to Fig. 8.

[0105] According to the third embodiment, when a protection target area including a protection target, such as a face or a license plate, whose size is larger than a certain value is present in an image, masking processing is performed on all protection target areas, and masking processing is not performed in other cases. For example, masking processing can be performed when the area of ​​the protection target is large compared to the area of ​​the block, and masking processing is not performed in other cases. This makes it possible to prevent unnecessary masking processing, reduce the amount of image data, and remove personal information.

[0106] <Embodiment 4> In the third embodiment, when a protection target area including a large-sized protection target exists in an image, the masking process is performed on all protection target areas in the image. In contrast, in the fourth embodiment, the masking process is performed only on the protection target area including the large-sized protection target. The configurations of the image processing system 100 and the server 2 are the same as those in the first embodiment. The configuration of the image compression device 1 is the same as that of the second embodiment, except for the configuration of the masking processing unit 15, which is partially different.

[0107] 10, masking processing unit 15 refers to the extraction results of the protection target area and protection target size acquired from protection target area extraction unit 11, and determines whether or not there is a protection target area including a large-sized protection target in the image captured by camera 1A. For example, according to the extraction results of the protection target area and protection target size shown in FIG. 12, the protection target flag of block number 80 (a block included in protection target area 41A) is "1" and the protection target size is "large." Therefore, there is a protection target area including a large-sized protection target in the image.

[0108] The masking processor 15 performs masking to remove personal information from a protection target area including a large protection target, but does not perform masking on other areas (protection target areas including a small protection target and areas other than the protection target area).

[0109] FIG. 14 is a flowchart showing an example of the procedure of the process executed by the camera 1A and the image compression device 1. Camera 1A executes the process of step S101, and image compression device 1 executes the process of step S102. The processes of steps S101 and S102 are the same as the processes of steps S101 and S102 described with reference to FIG.

[0110] Based on the result of extraction of the protection target area, the masking processing unit 15 determines whether or not the image captured by the camera 1A includes a protection target area that includes a large-sized protection target (step S121).

[0111] If a protection target area including a large-sized protection target is included (YES in step S121), the masking processing unit 15 executes masking processing to remove personal information from the protection target area including the large-sized protection target (step S131).

[0112] If a protection target area including a large-sized protection target is not included (NO in step S121), the process of step S131 is not executed.

[0113] Thereafter, the image compression device 1 executes the processes of steps S112, S104, and S105. The process of step S112 is the same as the process of step S112 described with reference to Fig. 11. The processes of steps S104 and S105 are the same as the processes of steps S104 and S105 described with reference to Fig. 8.

[0114] For example, if the size of the protection target included in the protection target area is large, such as a face or license plate, there is a possibility that personal information may be read from the decompressed image. In contrast, according to the fourth embodiment, by performing a masking process on the protection target area including the large protection target, it is possible to reliably remove personal information and reduce the amount of image data.

[0115] <Embodiment 5> In the second to fourth embodiments, the image compression device 1 performs the masking process on the protection target area. Alternatively, the server 2 may perform the masking process on the protection target area.

[0116] The configurations of the image processing system 100 and the image compression device 1 are the same as those in the first embodiment. However, the protection target area extraction unit 11 of the image compression device 1 is the same as that shown in embodiment 3. That is, the protection target area extraction unit 11 extracts the protection target area and the size of the protection target included in the protection target area from the image captured by the camera 1A.

[0117] FIG. 15 is a block diagram showing an example of the configuration of the server 2 according to the fifth embodiment. The server 2 includes an image receiving unit 21, an image decompression unit 22, an image processing unit 23, a display control unit 24, a storage device 25, and a masking processing unit 26.

[0118] The server 2 may be configured as a computer system including a processor such as a CPU and a storage device 25 such as a volatile memory or a non-volatile memory. Each of the processing units 21, 22, 23, 24, and 26 performs its function by executing a computer program stored in the storage device 25 on the processor.

[0119] The image receiving unit 21 and the image decompression unit 22 are the same as those in the first embodiment, except that the image decompression unit 22 outputs the position information of the protection target area acquired from the image receiving unit 21 to the masking processing unit 26.

[0120] The masking processing unit 26 performs masking processing on the protection target area including the large-sized protection target in the decompressed image based on the position information of the protection target area. The masking processing unit 26 outputs the image after the masking processing to the image processing unit 23. An example of the masking processing is as shown in the second embodiment. Masking processing section 26 outputs the image after the masking process to the image processing unit 23.

[0121] The image processing unit 23 performs predetermined image processing on the image after the masking processing. Examples of the image processing are as shown in the first embodiment.

[0122] The image processing unit 23 outputs the result of the image processing to the display control unit 24. The image processing unit 23 also writes the result of the image processing into the storage device 25. The display control unit 24 displays the image processing results obtained from the image processing unit 23 on the display device 2A.

[0123] FIG. 16 is a flowchart illustrating an example of a procedure of a process executed by the server 2. The image receiving unit 21 receives the compressed image to which the position information of the area to be protected has been added from the image compression device 1 via the network 3 (step S201). An example of the position information is as shown in FIG.

[0124] The image decompression unit 22 decompresses the compressed image received by the image receiving unit 21 to restore the original image (step S202).

[0125] The masking processing unit 26 determines whether or not a protection target area including a large-sized protection target exists in the decompressed image based on the position information of the protection target area (step S211).

[0126] If a protection target area including a large-sized protection target is included (YES in step S211), the masking processing unit 26 executes masking processing to remove personal information from the protection target area including the large-sized protection target (step S212).

[0127] If a protection target area including a large-sized protection target is not included (NO in step S211), the process of step S212 is not executed.

[0128] The image processing unit 23 performs predetermined image processing on the image after the masking processing (step S203). The image processing unit 23 writes the result of the image processing into the storage device 25 (step S204).

[0129] The display control unit 24 acquires the result of the image processing from the image processing unit 23 and causes the display device 2A to display it (step S205).

[0130] According to the fifth embodiment, the server 2, which has acquired the compressed image and the location information of the protection target area, can decompress the image while changing the compression rate based on the location information. Furthermore, if the protection target area has not been masked, the server 2 can mask the protection target area based on the location information. This allows the server 2 to perform other image processing after reliably removing personal information.

[0131] <Variation 1> In the third and fourth embodiments, masking processing is performed on the protection target area based on the size of the protection target included in the protection target area, but the compression rate when compressing the protection target area may also be changed based on the size of the protection target included in the protection target area.

[0132] For example, the image compression device 1 may set a compression rate for a protection target area including a large-sized protection target higher than that for a protection target area including a small-sized protection target. In other words, the compression rates are highest in the order of the compression rate for a protection target area including a large-sized protection target, the compression rate for a protection target area including a small-sized protection target, and the compression rate for other areas.

[0133] For example, the larger the size of the protection target, such as a face or license plate included in the protection target area, the higher the compression rate can be, which makes it possible to reduce the amount of image data while reliably removing personal information.

[0134] <Variation 2> In the first to fifth embodiments, the protection target area extraction unit 11 of the image compression device 1 extracts the protection target area in units of blocks, but the extraction method is not limited to this.

[0135] For example, the protection target area extraction unit 11 may detect a protection target from an image and extract the protection target area by dividing the protection target area from other areas. Alternatively, the protection target area extraction unit 11 may detect a protection target from an image and extract a circumscribing rectangle of the protection target as the protection target area.

[0136] [Note] Some or all of the components constituting each of the above devices may be configured from one or more semiconductor devices such as system LSIs (Large Scale Integration).

[0137] The computer program described above may be distributed by recording it on a computer-readable non-transitory recording medium, such as a hard disk drive (HDD), a CD-ROM, a semiconductor memory, etc. The computer program may also be distributed by transmitting it via a telecommunications line, a wireless or wired communication line, a network such as the Internet, data broadcasting, etc. Furthermore, each of the above devices may be realized by a plurality of computers or a plurality of processors.

[0138] Furthermore, some or all of the functions of each of the above-described devices may be provided by cloud computing, that is, some or all of the functions of each device may be realized by a cloud server. Furthermore, at least some of the above-described embodiments and modifications may be combined in any manner.

[0139] The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present disclosure is defined by the claims, not by the above meaning, and is intended to include all modifications within the meaning and scope of the claims. [Explanation of symbols]

[0140] 1. Image compression device 1A Camera 2 Server 2A display device 3 Network 11 Protection target area extraction unit (extraction unit) 12 Area Compression Section (Compression Section) 13 Image transmission unit (output unit) 14 Storage device 15 Masking processing section 21 Image receiving unit 22 Image decompression unit 23 Image processing section 24 Display control unit 25 Storage device 26 Masking processing section 41 People 41A Protected Areas 42 People 43 People 43A Protected Areas 44 People 44A Protected Areas 45 People 45A Protected Area 100 Image Processing System

Claims

1. an extraction unit that extracts a protection target area from an image, the protection target area being an area including personal information to be protected; a masking processing unit that performs a masking process to remove personal information from the protection target area; a compression unit that compresses the image after the masking process by compressing the protection target area at a higher compression rate than an area other than the protection target area in the image.

2. The image compression device according to claim 1 , wherein the extraction unit extracts the area to be protected by determining whether each of a plurality of blocks obtained by dividing the image corresponds to the area to be protected.

3. The image compression device according to claim 1 , wherein the masking processing unit performs the masking process on the protection target area based on a size of the protection target included in the protection target area.

4. The image compression device according to claim 1 , wherein the compression section determines a compression rate for compressing the protection target area based on a size of the protection target included in the protection target area.

5. The image compression device according to claim 3 , wherein the size of the protection target is determined based on an area of ​​the protection target included in the protection target region and an area of ​​each block obtained by dividing the image.

6. The image compression device according to claim 1 , further comprising an output unit that outputs position information of the protection target area in the image.

7. extracting a protection target area from the image, the protection target area being an area including personal information protection targets; a step of performing a masking process to remove personal information from the protected area; compressing the image after the masking process by compressing the region to be protected at a higher compression rate than regions other than the region to be protected in the image.

Images