Ambiguous processing inspection device, ambiguous processing inspection method, and ambiguous processing inspection program

Abstract

To provide an ambiguous processing inspection device that can improve the efficiency of operations.SOLUTION: An ambiguous processing inspection device of the present invention has: an image acquisition unit that acquires a provided image; a specific part detection unit that detects whether a specific part of the body is included in the acquired provided image; an ambiguous processing inspection unit that determines the presence or absence of necessity of ambiguous processing treatment on a specific part area; and a registration unit that registers a processed image subjected to the ambiguous processing treatment, and a non-detected image in which the specific part is not detected, as treated images.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention is suitably applicable to, for example, a video portal site that posts moving images for viewers to watch.

Background Art

[0002] In recent years, detection technologies that use deep learning technology to detect an object to be identified with high accuracy from a still image have been spreading (see, for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] By the way, if a detection technology with such a configuration is used, it becomes possible to detect a specific part with high accuracy. Therefore, it is considered that a substantially complete inspection of specific parts by AI (Artificial Intelligence) becomes possible and the efficiency of the work is improved.

[0005] The present invention has been made to solve such problems, and an object thereof is to provide an ambiguous processing inspection device, an ambiguous processing inspection device, and an ambiguous processing inspection program that can improve the efficiency of work.

Means for Solving the Problems

[0006] To solve such problems, in the ambiguous processing inspection device of the present invention, an image acquisition unit that acquires a provided image, a specific part detection unit that detects whether or not a specific part of the body is included in the acquired provided image, an ambiguous processing inspection unit that determines whether or not there is a need for the ambiguous processing for a specific part region including the specific part, An ambiguous processing inspection apparatus characterized by comprising

[0007] Also, in the ambiguous processing inspection method of the present invention, a specific part detection step of detecting whether a specific part of the body is included in the acquired provided image, and an ambiguous processing inspection step of determining whether there is a need for ambiguous processing for a specific part region including the specific part, An ambiguous processing inspection method characterized by comprising the above.

[0008] For a computer, a specific part detection step of detecting whether a specific part of the body is included in the acquired provided image, and an ambiguous processing inspection step of determining whether there is a need for the above-mentioned ambiguous processing for a specific part region including the specific part, An ambiguous processing inspection program characterized by causing the above to be executed.

Effect of the Invention

[0009] The present invention can realize an ambiguous processing inspection apparatus, an ambiguous processing inspection method, and an ambiguous processing inspection program that can improve the efficiency of operations.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Embodiments for Carrying Out the Invention

[0011] Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings.

[0012] <Embodiment> As shown in FIG. 1, 1 shows a video posting system as a whole. The video posting system 1 includes a posting terminal 3 that shoots a subject SJ and performs user operation input, a viewer terminal 4 that receives and views a video, and a server 2 that performs video distribution. Communication is executed between each device (posting terminal 3, viewer terminal 4, and server 2) using various telecommunication circuits such as the Internet using wireless and wired means.

[0013] As shown in FIG. 2, the server 2 has a computer configuration, and a control unit 21 composed of a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory) comprehensively controls the entire server 2, and according to a mosaic processing program and a video distribution program stored in a storage unit 25 in advance, it is configured to execute mosaic inspection and processing and video distribution processing. The server 2 communicates with other devices via an external interface 33 and executes mosaic inspection and processing.

[0014] As shown in FIG. 3, the contributor terminal 3 also has a computer configuration. A control unit 21 composed of a CPU, a ROM, and a RAM controls the whole in an overall manner and executes various processes according to a program stored in a storage unit 25 in advance. The contributor terminal 3 has an external interface 33, a display unit 34, an operation input unit 36, and a camera 39. In response to an operation on the operation input unit 36 by the user, the camera 39 performs shooting, the captured image data is displayed on the display unit 24, and the imaging data is stored in the storage unit 35. This shooting data is appropriately edited and supplied to an external device by communication via the external interface 33 according to the user's operation.

[0015] Although not shown, the viewer terminal 4 also has a computer configuration. A control unit 41 composed of a CPU, a ROM, and a RAM controls the whole in an overall manner and executes video viewing processing using video viewing application software (hereinafter abbreviated as an app) stored in a storage unit 45 in advance. The viewer terminal 4 has an external interface 43, a display unit 44, and an operation input unit 46. In response to an operation on the operation input unit 46 by the user (viewer), communication is performed via the external interface 43, and the supplied image data is displayed on the display unit 44.

[0016] In the video posting system 1 shown in FIG. 1, the case where the subject SJ is photographed by the camera 39 of the contributor terminal 3 and distributed to the viewer terminal 4 via the server 2 is shown.

[0017] Specifically, the posting user activates the posting application via the posting terminal 3 and starts the video posting process. This posting application is a so-called browser application that operates on the built-in browser. The posting user first creates an account for the posting application by specifying a user ID and password. At this time, the user may be authenticated by social login using the user's SNS (Social Networking Service) account by specifying the user's SNS. In this case, the user ID and password in the posting application are the same as the user ID and password in the SNS account.

[0018] When the server 2 is supplied with the user ID and password from the posting terminal 3, it authenticates the posting user and supplies the image data of the user's personal page screen (not shown) corresponding to the authenticated posting user to the posting terminal 3. The control unit 31 of the posting terminal 3 causes the display unit 34 to display the personal page screen. A posting button labeled "Post" is displayed on the personal page screen. By specifying the image data and operating the posting button, the specified image data is supplied to the server 2.

[0019] When the image data is supplied, the control unit 21 of the server 2 executes a mosaic inspection and processing operation (described in detail later) on the supplied image data (hereinafter referred to as the provided image data), and stores the processed image data in the storage unit 25 in association with the user ID. The execution of the mosaic inspection and processing operation does not have to be immediately after the supply of the provided image data. For example, the provided image data may be temporarily stored in the storage unit 25, and the mosaic inspection and processing operation may be executed when the server 2 is idle, such as at midnight or early morning.

[0020] A posting user database is stored in the storage unit 25 of the server 2, and payment information such as a bank account or an application account of a payment application is registered in association with the user ID. That is, the posting user database stores the user ID, password, and payment information. Also, the image ID of the processed image data is registered in association with the user ID.

[0021] Also, the browsing user who performs the browsing activates the browser via the viewer terminal 4, accesses the video browsing site, and starts the video browsing process. First, the browsing user creates an account on the browsing site by specifying a user ID and a password, and registers payment information for performing payment for the browsing, such as a credit card or a payment application. As a result, the user ID, password, and payment information are registered in the viewer database stored in the storage unit 25 of the server 2.

[0022] When the image ID of the image to be browsed and the user ID of the viewer are supplied from the viewer terminal 4 in response to the operation of the browsing user, the control unit 21 of the server 2 determines whether the browsing user has the browsing authority based on the payment management database stored in the storage unit 25.

[0023] For example, when charging a monthly fee for each poster, it is confirmed whether the browsing user has completed the payment process for the charge to the poster this month. Also, when charging each time for each image, it is confirmed whether the payment process for the image has been completed. When charging a monthly fee for each site, it is confirmed whether the browsing user has completed the payment process for the charge this month.

[0024] When the browsing user has not completed the payment process, the control unit 21 of the server 2 guides the viewer terminal 4 to the payment process and induces it to the payment process. When the payment process is completed, the control unit 21 grants the browsing authority to the browsing user by performing processes such as setting a flag in the corresponding charge column in the user ID of the viewer in the payment management database or registering the charged image ID.

[0025] The control unit 21 of the server 2 supplies the processed image corresponding to the supplied image ID to the browsing user who has the browsing authority. As a result, the processed image is displayed on the display unit 44 of the viewer terminal 4.

[0026] As a general business practice, blurring is considered insufficient as an appropriate obscuring process for hiding sexual expressions, and mosaic processing that makes six or more pixels the same color is often required. This standard is, so to speak, the safety line for legal compliance.

[0027] As described above, in the video posting system 1, it is assumed that many posting users will post videos, and it is assumed that not only businesses but also individuals will post. In the case of the video posting system 1 with such a large number of posts, there may be cases where appropriate obscuring processing is not performed on parts (specific parts of the body such as genitals and anus) that should be mosaicked appropriately to ensure legal compliance, such as mosaic processing of six or more pixels. Note that appropriate obscuring processing means obscuring processing that is considered not problematic in practice and can ensure legal compliance. Hereinafter, the case where appropriate obscuring processing is mosaic processing will be described.

[0028] On the other hand, in such a video posting system 1, instead of carefully edited images, many people with various completions are required to make diverse posts, and there are a very large number of provided images to be posted.

[0029] Therefore, in the present invention, after detecting most images including specific parts in the provided images that have been posted, the degree of mosaic processing in the detection area of the specific part (hereinafter referred to as the specific part area) is determined, and appropriate mosaic processing is executed as necessary.

[0030] First, the control unit 21 of the server 2 detects whether or not the provided image includes a specific part. As shown in FIG. 4, the specific part detection unit 21A (see FIG. 2) of the control unit 21 generates a downscaled image (d1 to d10 in the figure) by downscaling the frame image (f1 to f10 in the figure) of the provided image. As the downscaling, one or more of resolution reduction, color number reduction, grayscale conversion, etc. are appropriately selected and preferably combined as necessary.

[0031] The specific part detection unit 21A compares the color arrays of each pixel between the downsampled images. Since the differences in the color arrays are small (the number of differences is less than the threshold), it groups the frame images for each similar group (X, Y, Z in the figure), and selects one unique frame for each group. There is no limitation on the position of this unique frame, and it is selected from preset positions such as the beginning, middle, or end of the group, for example.

[0032] The specific part detection unit 21A determines whether a specific part is included in the unique frame.

[0033] Here, the detection of the specific part is performed using AI (Artificial Intelligence) technology using deep learning. Since the specific part is assumed to be reflected in various situations, various situations (angles, postures, skin color, etc.) are learned as teacher data.

[0034] Also, the cut-off threshold in the detection is set low, and since the frame is further downsized, almost all specific parts included in the unique frame are detected as specific part regions regardless of the presence or absence of mosaic processing in the provided image. Note that almost all specific parts means 80% or more, more preferably 90% or more. Therefore, almost all specific parts are detected as specific part regions, except for those that have been processed sufficiently so as not to be recognized as specific parts already.

[0035] The mosaic inspection unit 21B of the control unit 21 calculates the degree of mosaic processing in the detected specific part (hereinafter referred to as this specific part region). The degree of mosaic processing is calculated from the degree of change in the color array of the pixels in the specific part region.

[0036] Specifically, based on the degree of change in pixel color intensity, such as the ratio of edge pixels with a large change in color intensity compared to adjacent pixels (the smaller the ratio, the smaller the pixelation), the frequency of color intensity (the smaller the frequency, the smaller the dispersion and the smaller the pixelation), and the standard dispersion of color intensity (the larger the standard dispersion, the smaller the pixelation), the degree of mosaic processing (the magnitude of pixelation) in a specific site area is calculated.

[0037] When the degree of mosaic processing is large, there is no need for further mosaic processing. Therefore, the unique frame is used as the processed image, and for the non-detected image (frame image) corresponding to the unique frame and similar frames belonging to the group of unique frames, it is registered as the processed image through processing such as setting a flag.

[0038] On the other hand, when the degree of mosaic processing is small, it is determined that mosaic processing needs to be performed because no mosaic processing has been done or it is insufficient (the number of mosaic pixels is less than 6).

[0039] Based on the position and range of the specific site area in the unique frame, which is a down-frame, the mosaic area determination unit 21C determines the mosaic area to be subjected to mosaic processing. That is, as shown in FIG. 6, the mosaic area determination unit 21C determines the area corresponding to the specific site area (dashed line) as the mosaic area (solid line) in the frame image (before size reduction / see FIG. 6(B)) corresponding to the unique frame (see FIG. 6(A)).

[0040] At this time, the mosaic area determination unit 21C determines the mosaic area by adding α (α is an integer of 0 or more, preferably a natural number of 1 or more / α = 2 in FIG. 6) to the pixels corresponding to the specific site area, considering the pixels that may have been cut off during downsampling although the specific site is actually included. When the specific site area is detected to include the peripheral pixels of the specific site, α may be 0.

[0041] The mosaic processing unit 21D of the control unit 21 executes a mosaic processing using a calculation formula according to a predetermined formula on the mosaic area, and registers the processed image (frame image) corresponding to the unique frame as a processed image.

[0042] Furthermore, for the frame images corresponding to the similar frames belonging to the group of unique frames, the mosaic processing unit 21D determines the same coordinates as the mosaic area corresponding to the unique frame as the mosaic area, executes the mosaic processing to generate a processed image, and registers the processed image (frame image) corresponding to the similar frame as a processed image.

[0043] At this time, the mosaic processing unit 21D may perform the mosaic processing by directly using the result of the mosaic processing on the frame image corresponding to the unique frame and pasting it on the mosaic area of the frame image corresponding to the similar frame, or may calculate the mosaic area and perform the mosaic processing.

[0044] In this way, by detecting almost all of the images including the specific part in the frame image and judging the appropriateness of the mosaic processing, in the video posting system 1 where many ordinary people post images, it is possible to perform a full inspection on whether the mosaic processing is appropriately performed, and improve the legal compliance safety of the video posting system 1.

[0045] Next, the mosaic processing procedure executed according to the mosaic inspection program will be described using the flowchart of FIG. 8.

[0046] When the control unit 21 of the server 2 starts the mosaic inspection and processing, it enters from the start step and moves to step S1. In step S1, the control unit 21 downsizes the frame image in the provided image to generate a down frame, and moves to the next step S2.

[0047] In step S2, when the control unit 21 determines the similarity between down frames, groups them for each similar frame, and selects one unique frame for each group, it proceeds to the next step S3.

[0048] In step S3, the control unit 21 determines whether a specific part is included in the unique frame. If a negative result is obtained here, this indicates that there is no need for mosaic processing. At this time, the control unit 21 proceeds to step S8.

[0049] On the other hand, if an affirmative result is obtained in step S3, this indicates that it is necessary to inspect the mosaic processing. At this time, the control unit 21 proceeds to the next step S4.

[0050] In step S4, the control unit 21 inspects the degree of mosaic processing in the specific part area and determines whether the mosaic processing is insufficient. If a negative result is obtained here, it indicates that the mosaic processing is sufficient and there is no need for mosaic processing. At this time, the control unit 21 proceeds to the next step S7.

[0051] On the other hand, if an affirmative result is obtained in step S4, this indicates that there is a need for mosaic processing and it proceeds to the next step S5.

[0052] In step S5, the control unit 21 determines the mosaic area in the frame image corresponding to the specific part area of the unique frame and the frame images corresponding to the similar frames in the same group based on the corresponding one and range, and proceeds to the next step S6.

[0053] In step S6, the control unit 21 performs a mosaic processing on the mosaic area and proceeds to the next step S7.

[0054] In step S7, the control unit 21 registers, as processed images, the frame images corresponding to the unique claims and similar frames for which the mosaic processing has been executed or for which it has been determined that the mosaic processing is unnecessary, and proceeds to the next step S9.

[0055] In step S9, the control unit 21 determines whether processing has been completed for all the frame images. If a negative result is obtained, this indicates that there are still unprocessed frame images. At this time, the control unit 21 returns to step S1 and continues the mosaic inspection and processing.

[0056] On the other hand, if an affirmative result is obtained in step S9, since processing for all the frame images has been completed, the control unit 21 proceeds to the end step to end the processing.

[0057] <Second Embodiment> Next, the second embodiment will be described with reference to FIGS. 9 to 11. In the second embodiment, the method for determining whether to perform the mosaic process is different from that of the first embodiment. In the second embodiment, a sign obtained by adding 100 to the corresponding part in the first embodiment is attached, and the same sign is attached to the same part, and the description of the same part as in the first embodiment is omitted.

[0058] In the second embodiment, a pre-mosaicked already ambiguous region and a specific part region are detected in parallel, and after comparing the two, a mosaic region to be mosaicked is determined.

[0059] Specifically, the specific part detection unit 21A of the control unit 121 (see FIG. 9) of the server 102 detects a specific part region from the unique frame. Also, the already ambiguous detection unit 121E detects an already ambiguous region from the unique frame.

[0060] When the mosaic area determination unit 121C does not detect a specific part area and an already ambiguous area (hereinafter referred to as these detected areas) from the unique frame, it determines that mosaic processing is unnecessary. When the mosaic area determination unit 121C detects one detected area from one unique frame, it determines the one detected area as the mosaic area.

[0061] Also, when the mosaic area determination unit 121C detects two or more detected areas from one unique frame, it determines which of the two or more detected areas should be the mosaic area for which mosaic processing should be performed.

[0062] When there is no area where pixels overlap, the mosaic area determination unit 121C determines all of the detected detected areas as the mosaic area.

[0063] On the other hand, when there is an area where the detected detected areas overlap in pixels, it is determined whether the two detected areas can be regarded as the same.

[0064] As shown in FIG. 10(A), when the pixels of one detected area are all included in the other detected area, the larger detected area is determined as the mosaic area.

[0065] Also, as shown in FIGS. 10(B) and 10(C), when only some of the pixels overlap among two or more detected areas, it is determined whether the two or more detected areas can be regarded as the same.

[0066] Based on the larger of the two detected areas, the mosaic area determination unit 121C determines that the two are not the same when the overlapping pixels that overlap are less than a predetermined overlap threshold (for example, 65 to 80%), and determines both of the two detected areas as the mosaic area.

[0067] For example, as shown in FIG. 10(B), the mosaic area determination unit 121C determines that the two are the same when the overlapping pixels that overlap are equal to or greater than the overlap threshold, based on the larger of the two detected areas.

[0068] At this time, the mosaic area determination unit 121C selects the detection area with the larger size and determines it as the mosaic area.

[0069] Also, FIG. 10(C) shows a case where two mosaic areas and one specific part area are detected. In this case, in the small mosaic area and the specific part area, since the overlapping pixels are less than the overlap threshold, the mosaic area determination unit 121C determines them as separate detection areas. Also, in the large mosaic area and the specific area, since the overlapping pixels are equal to or more than the overlap threshold, they are regarded as the same.

[0070] At this time, the mosaic area determination unit 121C determines the smaller mosaic area determined as separate and the larger one (the large mosaic area in the figure) of the large mosaic area and the specific area as the mosaic area.

[0071] In the second embodiment, by determining the identity of two or more detection areas and selecting the larger size as the mosaic area when they are determined to be the same, the certainty of the mosaic process for the specific part is enhanced while preventing the deterioration of the image quality due to excessive mosaic of the peripheral image of the area.

[0072] Next, the mosaic inspection and processing procedure executed according to the mosaic inspection program will be described with reference to the flowchart RT21 of FIG. 11.

[0073] When the control unit 121 of the server 102 starts the mosaic process, it moves from the start step to step S1. After downsizing the size of the frame image in steps S1 and S2, it extracts the unique frame and moves to the next step S21.

[0074] In step S21, the control unit 121 detects the specific part area from the unique frame and moves to the next step S22.

[0075] In step S22, the control unit 121 detects the already ambiguous area from the unique frame and proceeds to the next step S23. Note that steps S21 and S22 are processes that are performed in parallel, and the order may be swapped.

[0076] In step S23, the control unit 121 executes the selection of the detection area by overlapping determination for the detection area and proceeds to the next step S24.

[0077] In step S24, when it is determined that mosaic processing is necessary for the detection area, the process proceeds to the next step S25. In step S25, the control unit 121 determines the mosaic area for which mosaic processing is to be performed and proceeds to the next step S6.

[0078] In steps S6 to S7, the control unit 121 executes mosaic processing on the mosaic area and proceeds to the next step S9.

[0079] In step S9, the control unit 121 determines whether processing has been completed for all frame images. If a negative result is obtained, this indicates that there are remaining frame images that have not been processed. At this time, the control unit 121 returns to step S1 and continues the mosaic inspection and processing.

[0080] On the other hand, in step S9, if an affirmative result is obtained, since processing for all frame images has been completed, the control unit 121 proceeds to the end step and ends the processing.

[0081] <Operation and Effect> Hereinafter, the features of the invention group extracted from the above-described embodiments will be described while showing effects as necessary. In the following, for ease of understanding, the corresponding configurations in the above embodiments are appropriately shown in parentheses, etc., but are not limited to the specific configurations shown in the parentheses, etc. Also, the meanings and examples of the terms described for each feature may be applied as the meanings and examples of the terms described for other features described in the same wording.

[0082] In the above configuration, the blurred processing inspection apparatus (server 2) of the present invention includes an image acquisition unit (external interface 23) that acquires a provided image, a specific part detection unit (specific part detection unit 21A) that detects whether or not a specific part of the body is included in the acquired provided image, and a blurred processing inspection unit (mosaic inspection unit 21B) that determines the necessity of the blurred processing for a specific part region including the specific part. It is characterized by comprising the above.

[0083] Thereby, the blurred processing inspection apparatus can detect the presence or absence of a specific part in the provided image, and when the specific part is detected, it can determine the necessity of the blurred processing, and can automatically inspect a frame image that requires the blurred processing. For this reason, it is possible to inform the provider of a provided image that has not been subjected to appropriate blurred processing of NG or to determine non - permission for public disclosure.

[0084] In the blurred processing inspection apparatus of the present invention, when a specific part is detected by the specific part detection unit, a blurred region determination unit (mosaic region determination unit 21C) that determines a blurred region by adding the specific part region and a peripheral pixel number that is an integer of 0 or more, a blurred processing unit (mosaic processing unit 21D) that executes the blurred processing on the blurred region when it is determined that the blurred processing is necessary, and a registration unit (storage unit 25) that registers the processed image obtained by the blurred processing and a non - detection image in which the specific part is not detected as a processed image. It is characterized by having the above.

[0085] Thereby, it is possible to automatically perform blurred processing on a provided image that has not been subjected to appropriate blurred processing, and to process it into a state suitable for public disclosure of the provided image.

[0086] In the blurred processing inspection apparatus of the present invention, the blurred processing inspection unit The ambiguity processing inspection apparatus according to claim 1, wherein the presence or absence of the necessity of the ambiguity processing is determined by calculating the degree of the ambiguity processing in the specific part area.

[0087] Accordingly, not only the presence or absence of the ambiguity processing for the specific part area but also the appropriateness can be determined, so that the provided image can be processed into an appropriate state for publication.

[0088] In the ambiguity processing inspection apparatus of the present invention, the ambiguity processing is a mosaic processing, the ambiguity processing unit determines the degree of the ambiguity processing based on the degree of change in the color intensity of the specific part area characterized by

[0089] Accordingly, the degree of the mosaic processing can be determined, and it can be determined whether appropriate mosaic processing necessary for compliance with the law is performed.

[0090] In the ambiguity processing inspection apparatus of the present invention, the storage unit registers the image acquired from the provider terminal that is the provider of the image in association with the provider identifier unique to the provider, registers settlement information indicating a settlement method in association with the viewer identifier unique to the viewer who views the image, an image supply unit that supplies the processed image to the viewer terminal that has supplied the image supply request signal when the image supply request signal is supplied from the viewer terminal owned by the viewer; a settlement processing unit that executes a settlement process corresponding to the supply of the processed image for the viewer according to the settlement information; and a payment processing unit that executes a payment process corresponding to the supply of the processed image for the provider in association with the provider identifier, characterized by

[0091] Thus, the present invention can be applied to a video posting site that provides provided images that are highly likely to have been inappropriately retouched by a large number of posters, including so-called laypersons who are not professionals, and the effects of the present invention can be fully utilized.

[0092] In the retouching inspection apparatus of the present invention, the specific part detection unit extracts one unique frame from similar frames determined to be similar based on the similarity between frame images in the provided image, determines the ambiguous region in the unique frame, the retouching processing unit performs the retouching process on the ambiguous region for all of the unique frame and similar frames similar to the unique frame.

[0093] Thereby, since the ambiguous region of the similar frames can be determined based on the unique frame, the number of affirmative processes can be significantly reduced, and the number of unprocessable cases can be reduced.

[0094] In the retouching inspection apparatus of the present invention, the specific part detection unit identifies the specific part region in a downscaled frame obtained by downsizing the frame image in the provided image, the retouching processing unit determines the ambiguous region based on the position and range of the corresponding specific part region in the frame image, and performs the retouching process on the ambiguous region in the frame image. Note that the corresponding position and range in the frame image refer to the position and range of the pixel numbers of the frame image obtained by multiplying the reciprocal of the magnification when downsizing by the pixel numbers of the downscaled frame.

[0095] Thereby, since the specific part region can be detected in the downscaled frame, the detection frequency of the specific part can be improved while reducing the number of unprocessable cases.

[0096] In the retouching inspection apparatus of the present invention, The specific part detection unit In the downsampled frame obtained by downsizing the frame image in the provided image, based on the similarity between frames, extracts one unique frame from the similar frames determined to be similar, and identifies the specific part area in the unique frame. The ambiguity processing unit In the frame image corresponding to the unique frame, determines the position and range of the ambiguous area, and performs the ambiguity processing on the ambiguous area in the frame image corresponding to the unique frame and the ambiguous area in the frame image corresponding to the similar frame similar to the unique frame. It is characterized by the above.

[0097] Accordingly, since the ambiguous area determined for the frame image corresponding to the unique frame can be directly applied to the frame image corresponding to the similar frame, the unprocessable situation can be significantly reduced.

[0098] In the ambiguity processing apparatus of the present invention, The ambiguity processing unit As the ambiguity processing for the ambiguous area in the similar frame, Copies the result of the ambiguity processing for the frame image corresponding to the unique frame and pastes it onto the ambiguous area in the identified frame image. It is characterized by the above.

[0099] Accordingly, since the ambiguity processing determined for the frame image corresponding to the unique frame can be directly applied to the frame image corresponding to the similar frame, the unprocessable situation can be significantly reduced.

[0100] In the ambiguity processing inspection apparatus of the present invention, In the provided image, it has an already ambiguous area detection unit that detects an already ambiguous area where ambiguity processing has already been performed. The ambiguity processing inspection unit Based on the specific part area and the already ambiguous area, it is characterized by determining the necessity of the blurring process.

[0101] Thereby, even if it is not detected as a specific part area, the already ambiguous area can be deliberately detected and judged, so that the detection accuracy of the specific part can be improved.

[0102] The blurring process inspection unit In the specific part area and the already ambiguous area, when the overlapping pixels that overlap are equal to or more than the overlap threshold determination, the specific part area and the already ambiguous area corresponding to the overlapping pixels are regarded as the same.

[0103] Thereby, while increasing the accuracy of mosaic detection, it is possible to prevent the mosaic process from being applied to the problem, and an appropriate mosaic process can be executed.

[0104] <Other embodiments> In the above-described embodiment, the case where mosaic processing of 6 pixels or more is performed as appropriate blurring has been described, but the present invention is not limited thereto. As long as the law can be complied with, when another technology has penetrated, the present invention can be applied to that technology. Further, the present invention can also be applied to still images.

[0105] Also, it is not always necessary to perform size reduction or extraction of unique frames, and inspection can be performed in the original size or all frames can be fully inspected.

[0106] In the above-described embodiment, when the degree of blurring process is insufficient, the mosaic processing is directly executed, but the present invention is not limited to this. For example, it can also be used as an inspection device that only performs inspection and outputs an inspection result indicating a non-conforming frame image.

[0107] In the above-described second embodiment, the identity of the detection regions is determined, but the present invention is not limited to this. For example, it is also possible to determine that all of the detection regions are mosaic regions. In this case, the certainty of the mosaic process can be enhanced.

[0108] In the above-described embodiment, the case where the present invention is applied to a portal site for video posting has been described, but the present invention is not limited to this. For example, it is also possible to apply the present invention to the editing work of video movies sold individually by download or the like.

Industrial Applicability

[0109] The present invention can be applied to, for example, a posting site where videos can be easily posted simply by registering an account.

Explanation of Signs

[0110] 1: Video posting system 2: Server 3: Poster terminal 4: Viewer terminal 21: Control unit 21A: Specific part detection unit 21B: Mosaic inspection unit 21C: Mosaic region determination unit 21D: Mosaic processing unit 23: External interface 24: Display unit 25: Storage unit SJ: Subject

Claims

1. Detecting a specific part, including mosaic processing / uniform mosaic processing on the specific part An image acquisition unit that acquires a provided image; A specific part detection unit that detects whether a specific part of the body is included in the acquired provided image; An ambiguity processing inspection unit that determines whether there is a need for ambiguity processing on a specific part region having a specific part; An ambiguity processing inspection apparatus, characterized by comprising the above components.

2. B. Inspecting the necessity of ambiguity processing, only necessary when necessary (not necessary, weak) When it is determined that the ambiguity processing is necessary, When the specific part is detected by the specific part detection unit, an ambiguity region determination unit that specifies the specific part region and an ambiguity region obtained by adding a peripheral pixel number that is an integer of 0 or more; An ambiguity processing unit that performs the ambiguity processing on the ambiguity region; A registration unit that registers the processed image obtained by the ambiguity processing and the non-detection image in which the specific part is not detected as a processed image; The ambiguity processing inspection apparatus according to claim 1, characterized by comprising the above components.

3. B-2. Necessity = judged from the range and degree of the area of the ambiguity processed region The ambiguity processing inspection unit, Determines the necessity of the ambiguity processing by calculating the degree of the ambiguity processing in the specific part region; The ambiguity processing inspection apparatus according to claim 1 or claim 2, characterized by the above.

4. The ambiguity processing is A mosaic processing, The ambiguity processing unit, Determines the degree of the ambiguity processing based on the degree of change in the color intensity of the specific part region; The ambiguity processing inspection apparatus according to claim 3, characterized by the above.

5. A. Submission site The storage unit, Registers the image acquired from the provider terminal that is the provider of the image in association with the provider identifier unique to the provider, Registers settlement information indicating a settlement method in association with a viewer identification number unique to a viewer who views the image, An image supply unit that supplies the processed image to the viewer terminal that has supplied the image supply request signal when an image supply request signal is supplied from the viewer terminal owned by the viewer; A settlement processing unit that performs a settlement process corresponding to the supply of the processed image on the viewer according to the settlement information; A payment processing unit that performs a payment process corresponding to the supply of the processed image on the provider in association with the provider identifier; The ambiguity processing inspection apparatus according to claim 1 or claim 2, characterized by comprising the above components.

6. C. Extraction of unique frame The specific part detection unit, Based on the similarity between frame images in the provided image, one unique frame is extracted from the similar frames determined to be similar, and the ambiguous region is determined in the unique frame. The ambiguous processing unit is For all of the unique frame and all similar frames similar to the unique frame, the ambiguous processing is executed on the ambiguous region. The ambiguous processing inspection apparatus according to claim 1 or claim 2, characterized in that.

7. D. Size reduction The specific part detection unit is In the downsampled frame obtained by downsizing the frame image in the provided image, the specific part region is specified. The ambiguous processing unit is Based on the position and range of the corresponding specific part region in the frame image, an ambiguous region is determined, and the ambiguous processing is executed on the ambiguous region in the frame image. The ambiguous processing inspection apparatus according to claim 1 or claim 2, characterized in that.

8. C + D The specific part detection unit is In the downsampled frame obtained by downsizing the frame image in the provided image, based on the similarity between frames, one unique frame is extracted from the similar frames determined to be similar, and the specific part region is specified in the unique frame. The ambiguous processing unit is In the frame image corresponding to the unique frame, the position and range of the ambiguous region are determined, and the ambiguous processing is executed on the ambiguous region in the frame image corresponding to the unique frame and the ambiguous region in the frame image corresponding to the similar frame similar to the unique frame. The ambiguous processing inspection apparatus according to claim 1 or claim 2, characterized in that.

9. Mosaic copy, paste on similar frames The ambiguous processing unit is As the ambiguous processing on the ambiguous region in the similar frame, The result of the ambiguous processing on the frame image corresponding to the unique frame is replicated and pasted on the ambiguous region in the specified frame image. The ambiguous processing inspection apparatus according to claim 8, characterized in that.

10. Detection of already ambiguous region (second embodiment) In the provided image, it has an already ambiguous region detection unit that detects an already ambiguous region where ambiguous processing has already been performed. The ambiguous processing inspection unit is Based on the specific part region and the already ambiguous region, it discriminates whether there is a need for ambiguous processing. The blurred processing inspection apparatus according to claim 1, characterized in that...

11. Overlap determination The blurred processing inspection unit is configured to: In the specific part area and the already blurred area, when overlapping pixels that overlap are determined to be equal to or greater than the overlap threshold determination, the specific part area and the already blurred area corresponding to the overlapping pixels are regarded as the same. The blurred processing inspection apparatus according to claim 10, characterized in that...

12. A specific part detection step of detecting whether a specific part of the body is included in the acquired provided image, and A blurred processing inspection step of determining whether or not there is a need for the blurred processing for a specific part area including the specific part A blurred processing inspection method, characterized by comprising the above.

13. For a computer, A specific part detection step of detecting whether a specific part of the body is included in the acquired provided image, and A blurred processing inspection step of determining whether or not there is a need for the blurred processing for a specific part area including the specific part A blurred processing inspection program, characterized by causing the above to be executed.

Images