JP7124912B2 - Information processing device, information processing method, and information processing program - Google Patents

Description

The present invention relates to technology for identifying face images that include occluded areas.

Image recognition devices for identifying facial images are well known. In such an image recognition apparatus, recognition accuracy decreases when a part of the face in an input face image or a registered face image is blocked by an object such as glasses, sunglasses, or a mask.

One technique related to such problems is described in Japanese Unexamined Patent Application Publication No. 2002-200013. The related technology described in Patent Document 1 calculates an inter-pattern distance value between each of a plurality of small regions in an input image and each small region of a registered image at a position corresponding to each small region. Then, this related technique identifies an input image based on an integrated distance value calculated using only a predetermined number of inter-pattern distance values in ascending order. In this related technique, a small region having a large inter-pattern distance value between the input image and the registered image does not contribute to identification. Therefore, with this related technology, it is possible to avoid small areas in which the input image and the registered image are greatly different from each other due to the shading, and perform matching.

Another technique related to such problems is described in Japanese Unexamined Patent Application Publication No. 2002-200320. In the related technology described in Patent Literature 2, a plurality of pieces of face information are registered in advance for each user with different wearing states of wearing objects (such as eyeglasses). Then, this related technique identifies candidate face information similar to the input face image among the plurality of pieces of face information. Furthermore, this related technology identifies different type face information having a wearing state different from that of the candidate face information, among a plurality of pieces of face information registered for a user corresponding to the candidate face information. Then, this related technique determines that the input face image represents the user corresponding to the candidate face information when the input face image and the different type face information are similar in a region other than the wearing region of the wearable object. In this way, this related technology can perform matching regardless of the wearing state of the wearing object in the input face image.

In addition, another technique related to such problems is described in Japanese Patent Laid-Open No. 2002-200033. The related technology described in Patent Document 3 extracts a first feature amount from a region that avoids the wearable object region when it is determined that the wearable object is present in the face image. Then, in this related technique, a face image with a wearable object is registered together with the first feature amount. Further, in this related technique, when it is determined that there is no wearable item in the face image, in addition to the first feature amount, a second feature amount is extracted from an area including an assumed wearable item area. Then, in this related technique, a face image without a wearable object is registered together with the first feature amount and the second feature amount. Also, this related technique extracts a first feature amount or a second feature amount from an input face image, and performs matching with the first feature amount or the second feature amount of a registered image. In this way, this related technology performs matching even when a registered face image has a wearable object and an input face image does not have a wearable object, or even when a registered face image does not have a wearable object but an input face image has a wearable object. be able to.

In addition, another related technology related to such problems is described in Japanese Patent Application Laid-Open No. 2002-200042. The related technique described in Patent Document 4 divides an input face image and a reference face image into a plurality of small regions with feature points as vertices, and compares the corresponding small regions. Then, this related technique determines whether or not each small region of the input face image is a shielded region based on the comparison result.

In addition, another related technique related to such problems is described in Patent Document 5. The related technology described in Patent Literature 5 supplements a partially occluded region of an input face image with a pre-learned associative memory circuit. Then, this related technology performs matching using the complemented recalled image. In this way, this related technique performs matching assuming that all images are images without occluded regions.

Patent No. 4803214 JP 2014-115784 A Japanese Patent Application Laid-Open No. 2007-280250 Japanese Unexamined Patent Application Publication No. 2011-60038 JP 2007-148872 A

However, the related art described above has the following problems.

In the related technology described in Patent Literature 1, even when the input image and registered image have different faces, the placement and content of the shielded regions in each image may be similar. In such a case, this related technique ends up using a masked region with a small inter-pattern distance value for matching. Therefore, in this related technique, it may be determined that an input image and a registered image, which actually show different persons, show the same person.

Further, in the related technology described in Patent Literature 2, it is necessary to register a plurality of face images with different wearing states for each user. However, not all users can register such multiple face images. In addition, as the types of wearable objects increase, the user has to register face images with different wearing conditions for each wearable object. This creates a problem of convenience.

Further, in the related technology described in Patent Literature 3, as the types of assumed wearable objects increase, it becomes necessary to extract the first feature amount that avoids the assumed wearable object region for each wearable object. Therefore, in this related technique, it is necessary to extract the first feature amounts of different patterns by the number of types of wearables assumed, which increases the processing load. It should be noted that if the first feature amount is extracted from areas that avoid the wearable object areas of all assumed wearable items, even if the types of assumed wearable items increase, the number of patterns of the first feature amount does not increase. . However, in this case, the number of areas that can be used for matching among the areas representing the face decreases as the types of wearable objects increase. As a result, recognition accuracy decreases. In addition, as the number of possible types of wearable items increases, the cost of pre-learning the function of determining the presence or absence of the wearable items also increases.

Further, the related technique described in Patent Document 4 determines whether or not each small area of the input image is a shielded area based on the difference in brightness from that of the small area of the reference face image. However, due to the influence of illumination variations, some small areas may have a large difference in luminance from the reference face image even if they are not shielded areas. In this case, this related technique determines that a small area that is not a shielded area is a shielded area. In addition, a small region of the input face image may be shielded by a shielding object with little texture, and the corresponding small region of the reference face image may originally have little texture. At this time, the difference in luminance may be small between such small regions. In this case, this related technique determines that the small area that is the shielded area is not the shielded area. For example, there is a possibility that the difference in brightness between a part of the mask of a person wearing a white mask and the skin of the reference face is small. In this case, this related technique determines that the area blocked by the mask is not a blocked area. In this way, this related technique has a problem in accuracy of determination of the shielded area for each small area.

In addition, the related technology described in Patent Document 5 cannot cope with the case where the registered image has a shielded area. This is because learning for generating recall images that complement partially occluded regions requires registered images without occluded regions. Further, in this related technology, when an input face image representing a person other than the person represented by the registered image is input, a recalled image similar to the person represented by the registered image is generated. Therefore, this related technique may determine that such an input face image shows the same person as a registered face image that actually shows a different person.

The present invention has been made to solve the above problems. That is, an object of the present invention is to provide a technique for further improving the recognition accuracy of an image including a shielded area without increasing the cost of system construction and the processing load at the time of identification.

The information processing apparatus of the present invention includes a determining unit that determines a shielded area that is shielded in an input image that is an image representing a face, and an area other than an area that is associated with a shielding pattern based on the shielded area. an identification unit that identifies an image.

Further, the information processing method of the present invention determines a shielded region that is shielded in an input image that is an image representing a face, and uses regions other than regions associated with a shielding pattern based on the shielded region to determine the input image. identification.

Further, the information processing program of the present invention includes processing for determining a shielded region that is shielded in an input image that is an image representing a face; and a process of identifying an input image.

INDUSTRIAL APPLICABILITY The present invention can provide a technique for further improving the recognition accuracy of an image including an occluded area without increasing the cost of system construction and the processing load during identification.

1 is a functional block diagram of an image recognition device as a first embodiment of the present invention; FIG. 1 is a diagram showing an example of a hardware configuration of an image recognition device as a first embodiment of the invention; FIG. 4 is a flow chart for explaining the operation of the image recognition device as the first embodiment of the present invention; FIG. 5 is a functional block diagram of an image recognition device as a second embodiment of the present invention; FIG. 10 is a diagram schematically showing an example of small areas into which a face image is divided in the second embodiment of the present invention; FIG. FIG. 10 is a diagram schematically showing another example of small areas into which a face image is divided in the second embodiment of the present invention; FIG. 10 is a diagram schematically showing another example of small areas into which a face image is divided in the second embodiment of the present invention; FIG. 10 is a diagram schematically showing another example of small areas into which a face image is divided in the second embodiment of the present invention; 8 is a flow chart for explaining the operation of the image recognition device as the second embodiment of the present invention; FIG. 11 is a functional block diagram of an image recognition device as a third embodiment of the present invention; 13 is a flow chart for explaining the operation of the image recognition device as the third embodiment of the present invention; FIG. 11 is a functional block diagram of an image recognition device as a fourth embodiment of the present invention; 10 is a flow chart for explaining the operation of the image recognition device as the fourth embodiment of the present invention;

BEST MODE FOR CARRYING OUT THE INVENTION Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

(First embodiment)
FIG. 1 shows the configuration of an image recognition device 1 as a first embodiment of the invention. In FIG. 1 , the image recognition device 1 includes a determination section 11 and an identification section 12 .

Here, as shown in FIG. 2, the image recognition apparatus 1 includes a CPU (Central Processing Unit) 1001, a RAM (Random Access Memory) 1002, a ROM (Read Only Memory) 1003, and a storage device 1004 such as a hard disk. , an imaging device 1005, and an output device 1006. The ROM 1003 and the storage device 1004 store computer programs and various data for causing the computer device to function as the image recognition device 1 of this embodiment. The imaging device 1005 is a device such as a camera or a video camera that captures an image of a person's face to generate an image representing the face. The output device 1006 is a device such as a display that outputs information. The CPU 1001 loads computer programs and various data stored in the ROM 1003 and the storage device 1004 into the RAM 1002 and executes them, thereby controlling each part of the computer device.

In this case, the determination unit 11 is composed of an imaging device 1005 and a CPU 1001 that reads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. The identification unit 12 is composed of an output device 1006 and a CPU 1001 that loads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. The hardware configuration of the image recognition device 1 and its functional blocks is not limited to the configuration described above.

Next, each functional block of the image recognition device 1 will be described.

The determination unit 11 determines the masking pattern of the face in the face image by comparing the face image and the standard face image. Here, a face image is an image representing a face. For example, the determination unit 11 acquires a face image via the imaging device 1005 . A standard face image is an image representing a standard face. It is assumed that the standard face image is predetermined and stored in the storage device 1004, for example. Also, the face image and the standard face image are images obtained by cutting out a region representing the face from the corresponding image. Further, in the present embodiment, it is assumed that the face image and the standard face image are substantially the same in size, face orientation, and the like.

Further, the shielding pattern is information indicating which area is shielded by what kind of shield with respect to the face. Specifically, the shielding pattern may be information such as a mask covering the mouth and sunglasses covering the eyes. For example, the determination unit 11 compares the face image and the standard face image to estimate a masked region in which the face is masked by a masking object, and considers the distribution of the estimated masked region to determine the masking pattern. may

The identification unit 12 identifies the face image by excluding areas based on the shielding pattern in the face image. It is assumed that the exclusion area is determined in advance according to the assumed shielding pattern. Note that the excluded area does not necessarily have to match the shielded area estimated in the process of determining the shielding pattern by the determination unit 21 . Also, a known technology can be adopted for the facial image identification technology. The identification unit 12 then outputs the identification result to the output device 1006 .

The operation of the image recognition apparatus 1 configured as above will be described with reference to FIG.

First, the determination unit 11 compares an input face image (input face image) and a standard face image (step S1).

Next, the determination unit 11 determines the shielding pattern in the input face image based on the comparison result of step S1 (step S2).

Next, the identification unit 12 identifies the input face image by excluding the area based on the shielding pattern determined in step S2 in the input face image. Then, the identification unit 12 outputs the identification result (step S3).

With this, the image recognition apparatus 1 completes its operation.

Next, effects of the first embodiment of the present invention will be described.

The image recognition apparatus according to the first embodiment of the present invention can further improve recognition accuracy of an image including a shielded area without increasing the cost of system construction and the load of identification processing.

The reason for this is that the determination unit determines the masking pattern of the face image by comparing the face image and the standard face image, and the identification unit excludes the area based on the masking pattern in the face image, This is because identification is performed.

As described above, in the present embodiment, a face image is compared with a standard face image, and a masking pattern is determined in consideration of the distribution of masking regions in the face image. In addition, in the present embodiment, instead of excluding the masked region itself estimated to be masked by comparison with the standard face image, a masking pattern determined in consideration of the distribution of the masked region is determined. Identification is performed by excluding the region where the As a result, the present embodiment realizes shielding detection that is robust against a shielding area determination error. Further, according to the present embodiment, it is possible to prevent a decrease in authentication accuracy due to an error in determining a shielded area, and to improve the recognition accuracy of a face image.

Moreover, this embodiment does not require prior learning of the function for determining the presence or absence of a shield for each type of shield. In addition, this embodiment does not require pre-registration of multiple images corresponding to the presence or absence of shielding for each user. In addition, this embodiment does not require extraction of feature amounts for the number of types of shielding objects. In this embodiment, a standard face image is prepared in advance, and an area to be excluded is determined according to an assumed shielding pattern. do not have.

(Second embodiment)
Next, a second embodiment of the present invention will be described in detail with reference to the drawings. In each drawing referred to in the description of the present embodiment, the same reference numerals are assigned to the same configurations and steps that operate in the same manner as in the first embodiment of the present invention. Description is omitted.

First, FIG. 4 shows the configuration of an image recognition device 2 as a second embodiment of the present invention. In FIG. 4, the image recognition device 2 has a determination unit 21 in place of the determination unit 11 and an identification unit 22 in place of the identification unit 12, compared to the image recognition device 1 according to the first embodiment of the present invention. and further includes a dividing unit 23 and a reliability calculation unit 24 .

Here, the image recognition device 2 can be configured by hardware elements similar to those of the image recognition device 1 according to the first embodiment of the present invention described with reference to FIG. In this case, the dividing unit 23 is composed of an imaging device 1005 and a CPU 1001 that loads computer programs and various data stored in the ROM 1003 and the storage device 1004 into the RAM 1002 and executes them. The reliability calculation unit 24 is configured by the CPU 1001 that reads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. Note that the hardware configuration of the image recognition device 2 and its functional blocks is not limited to the configuration described above.

Next, each functional block of the image recognition device 2 will be described.

The dividing unit 23 divides the face image into small areas (also called patches). Note that the dividing unit 23 may acquire the face image via the imaging device 1005 . For example, the dividing unit 23 may divide the face image into small regions of the same shape. Further, the dividing unit 23 may divide the face image so that the shape of a certain small area is different from at least part of the other areas. Also, each divided small area may have a portion that overlaps with another small area. It is desirable that the dividing unit 23 divides the face image into small regions having the same arrangement and shape as the standard face image. For example, the dividing unit 23 may divide the standard face image into small regions in advance, and divide the input face image into similar small regions. Alternatively, the dividing unit 23 may acquire information of a standard face image divided into small areas in advance, and divide the face image into small areas similar to the acquired small areas.

Note that the dividing unit 23 may perform processing for normalizing the face image before dividing the face image. This is to cope with the case where the face image input to the image recognition device 2 is not normalized. Specifically, the dividing unit 23 generates a face image by normalizing the face in the face image. Note that the term "normalization" as used herein refers to deformation processing so that the angle, size, and the like of a face become a predetermined angle and size. For example, the dividing unit 23 may detect feature points such as the eyes and mouth in the input face image, and perform geometric transformation so that the feature points are positioned at predetermined positions. Further, when the image input to the image recognition device 2 is an image representing a wider area than the face, the dividing unit 23 may cut out the face image from the corresponding image and then perform normalization. For example, the division unit 23 may adopt the method disclosed in Japanese Patent No. 4653606 for the face image normalization process. Note that the dividing unit 23 may normalize the input face image using another method.

As a specific example, for example, the dividing unit 23 may divide the normalized face image into rectangular small regions of equal size in a grid pattern, as shown in FIG. Alternatively, as shown in FIG. 6, some or all of the subregions may be divided so as to overlap other subregions. In FIG. 6, the small area indicated by the thick solid-line rectangle overlaps with the four small areas indicated by the dashed-line rectangles. Alternatively, the dividing unit 23 may divide the normalized face image so that the size of a certain small area is different from the size of at least some of the other small areas, as shown in FIG. Alternatively, the dividing unit 23 may divide the normalized face image into subregions having shapes other than rectangles (for example, triangles) as shown in FIG. The division unit 23 may divide into small regions of arbitrary shape and size as long as the processing of the reliability calculation unit 24 and the identification unit 22 is not hindered.

The reliability calculation unit 24 compares the corresponding small regions of the face image and the standard face image. Then, the reliability calculation unit 24 calculates the reliability based on the comparison result for each small area of the face image. The reliability indicates the possibility of whether or not the small area is an occluded area where the face is hidden by an occluder. Note that the reliability may be smaller the higher the probability that the small area is a shielded area and larger the lower the probability. Conversely, the reliability may be higher when the small area is likely to be an occluded area and smaller when the possibility is lower.

A standard face image is stored in the storage device 1004 in advance. It is also assumed that the standard face image has been normalized in advance. For example, the standard face image may represent a face such as a general identification photograph that does not include the shielding area. Moreover, as the standard face image, it is desirable to adopt a face image that is easily collated with a larger number of face images. For example, as the standard face image, an image representing an average face obtained by averaging faces of a plurality of persons may be used. Also, as the standard face image, an image representing Wolf/Lamb (a sample whose degree of similarity to a plurality of people's faces satisfies a predetermined condition) may be used. Also, as the standard face image, an image representing an object (artificial object) artificially created to resemble the faces of a plurality of persons may be used. Also, the standard face image may be divided into small areas in advance. In this case, the subregions forming the standard face image may be divided by the dividing section 23 .

For example, the reliability calculation unit 24 may calculate the reliability by summing up differences in luminance values between corresponding small regions. However, reliability using luminance values is susceptible to illumination variations. Therefore, the reliability calculation unit 24 may calculate the reliability using the feature amount vector for each small area. The feature amount vector may be information obtained by feature extraction processing such as Gabor filter and LBP (Local Binary Pattern), which are often used in face authentication, for example. Alternatively, the feature amount vector may be information obtained by feature extraction processing such as a Haar-like filter that is often used in face detection. In this case, the reliability calculation unit 24 can calculate the reliability based on the distance or correlation between the feature quantity vectors between corresponding small regions. Based on such a feature amount vector, the reliability calculation unit 24 can calculate the reliability with reduced influence of illumination variation and the like. Note that when the reliability based on the distance between the feature amount vectors is calculated, the larger the reliability, the higher the possibility that the area is an occluded area. Also, when calculating the reliability based on the correlation between the feature amount vectors, the smaller the reliability, the higher the possibility that the area is an occluded area. For example, the reliability calculation unit 24 may calculate a normalized correlation value between feature amount vectors and use it as the reliability of each small region. In this case, it can be considered that the closer the reliability is to 1, the higher the possibility that the area is a non-shielded area, and the closer the reliability is to 0, the higher the possibility that the area is a shielded area. An example in which the reliability is a normalized correlation value will be mainly described below.

The determination unit 21 determines the masking pattern based on the distribution of reliability for each small area in the face image. Here, the determining unit 21 may determine whether or not the reliability distribution of the small area group forming the intermediate area satisfies the shielding condition for the intermediate area obtained by combining the small area groups within a predetermined range. . Then, when the shielding condition is satisfied, the determining unit 21 may estimate that the middle area is the shielding area. Then, the determination unit 21 may determine the shielding pattern based on the estimated distribution of the shielding areas.

For example, the determination unit 21 may determine whether the reliability of each small area forming the middle area is higher or lower than a threshold. Then, the determination unit 21 may regard the reliability of a small region having a discrimination result different from that of the surrounding small regions as noise. Specifically, when the determination unit 21 satisfies the shielding condition that a majority or more of the small regions in which the reliability of the group of small regions constituting the middle region is lower than the threshold, the middle region satisfies the shielding condition. may be estimated to be the occluded region. In addition, other conditions may be sufficient as shielding conditions. Then, the determination unit 21 may determine the shielding pattern based on the distribution of the medium areas estimated to be the shielding areas. Note that the middle region may be set stepwise. For example, the determination unit 21 may be configured such that a medium area obtained by grouping a group of small areas within a predetermined range is further grouped into a large area by a predetermined range. In such a case, the determination unit 21 may estimate that the large area is the shielded area when the distribution of the medium areas that constitute the large area that are estimated to be the shielded area satisfies the shielding condition. . In this case, the determination unit 21 may determine the shielding pattern based on the distribution of the large area estimated to be the shielding area. Further, the design is not limited to the three stages of the small area, medium area, and large area, and may be designed in multiple stages from the small area to the large area.

For example, the determination unit 21 may set the upper face region, which is a group of small regions above the nose in the face image, as the middle region. In this case, the determination unit 21 may determine that the masking pattern is due to sunglasses when the distribution of the reliability of the small region group in the upper face region satisfies the masking condition. Further, the determination unit 21 may set the lower face region, which is a group of small regions below the nose, as the middle region. In this case, the determination unit 21 may determine that the pattern is masked by a mask when the reliability distribution of the small region group in the lower face region satisfies the masking condition. In addition, the determination unit 21 may set the middle region according to various assumed shielding patterns.

Here, the advantage of determining the shielding pattern based on the distribution of the reliability of the small areas or the distribution of the shielding areas will be described. The reliability calculated for each small area tends to be unstable depending on the shooting conditions. Therefore, if the shielding determination is made based on the reliability for each small area, the possibility of an error is increased. Note that the error here refers to determining that a small area that is originally a shielded area is not a shielded area, or that a small area that is not originally a shielded area is determined to be a shielded area. Also, the reliability that can generate such an error can be said to be noise. If a non-shielded area is used for identification based on such shielding determination for each small area, there is a possibility that a small area (shielded area) that is originally not suitable for identification will be used for identification as a non-shielded area. be. Also, there is a possibility that a small area (area that is not a shielded area) that is originally suitable for an identification target may not be used for identification as a shielded area. Therefore, shielding determination based on the reliability of each small area reduces authentication accuracy.

On the other hand, masks and sunglasses are representative examples of face shielding patterns that often occur in daily life. These shielding patterns shield a very large area, on the order of 1/3 to 1/2 of the face. Considering such characteristics of occlusion patterns, if we assume that there is no noise in the reliability, there will be small regions with a reliability lower (or higher) than the threshold in the small region group with a higher (or lower) confidence than the threshold. It is unlikely that there will be a slight region. Therefore, if a reliability that is clearly different from the reliability of the surrounding small area is calculated, it can be considered that the reliability is noise.

Therefore, as described above, in consideration of such errors (noise), the determination unit 21 determines whether or not the distribution of the reliability in the middle region, which is a group of small regions within a predetermined range, satisfies the shielding condition. Then, it is estimated whether or not the middle area is a shielded area. As a result, the determination unit 21 can accurately estimate the shielded area compared to the case of determining whether or not each small area is a shielded area using only the reliability. As a result, the determination unit 21 can accurately determine the shielding pattern.

The identification unit 22 identifies each small area in the input face image except for the exclusion area based on the shielding pattern. It is assumed that the exclusion area is determined in advance according to the shielding pattern, as in the first embodiment of the present invention. For example, in the case of a shielding pattern by a mask, the area of the lower half of the face may be defined as an exclusion area. Also, if the pattern is shielded by sunglasses, the area of the upper half of the face may be defined as an exclusion area. Note that the exclusion area does not necessarily have to match the shielding area estimated in the process of shielding pattern determination by the determining unit 21 .

Then, the identification unit 22 identifies the face image based on the identification result for each small area other than the exclusion area. For example, the identification unit 22 may calculate the identification score for each small area other than the exclusion area based on the masking pattern in the face image. In that case, the identification unit 22 may calculate an integrated score by integrating the calculated identification scores for each small region, and output the integrated score as the identification result.

Specifically, the identification unit 22 compares each small area obtained by dividing the input face image with each small area constituting the registered face image, and performs identification by matching the corresponding small areas. A score may be calculated. In this case, the identification unit 22 compares corresponding small areas in areas that are not included in any of the exclusion areas based on the shielding patterns in the input face image and the registered image. Then, the identification unit 22 calculates an identification score for each small area other than the exclusion area based on the comparison result between the small areas. Then, the identification unit 22 integrates the identification scores to calculate an integrated score and outputs it as an identification result.

Here, it is assumed that the registered face images are stored in the storage device 1004 in advance. The storage device 1004 may store registered face images together with information representing shielding patterns of the registered face images. In addition, the storage device 1004 may store information representing small areas forming the registered face image together with the registered face image. The masking pattern of the registered face image may be information determined by the determination unit 21 when the registered face image is registered. Further, the information divided by the division unit 23 when the registered face image is registered may be used for the small areas that constitute the registered face image.

Further, the identification unit 22 may use a value based on the distance or correlation between the feature amount vectors as the identification score between the corresponding small regions. In this case, the feature amount vector of each small region may be information obtained by feature extraction processing such as Gabor fill or LBP, for example. However, it is desirable that the identification unit 22 uses a feature amount vector having a higher identification ability than the feature amount vector used when calculating the reliability with the standard face image for the process of identifying the registered face image. For example, the identification unit 22 may learn a conversion matrix for lower dimensions by performing linear discriminant analysis on feature amount vectors extracted from learning data. In this case, the training data may be facial images correctly labeled for each person. Such training data may be registered face images. For example, in a case where all the users of the image recognition device 2 can be identified (for example, in a case where it is used for entrance/exit management), it is desirable to use registered face images of all system users as learning data. Alternatively, the learning data is not limited to registered face images, and may be other labeled face images. For example, in a case where the image recognition device 2 is used for authentication of an unspecified number of people (for example, a case where it is used for suspicious person detection), as training data, regardless of whether the user is a system user or not, labeled It is desirable to use face images of a large number of persons. Then, the identification unit 22 generates a low-dimensional feature amount vector with enhanced identification ability by applying the aforementioned transformation matrix to the feature amount vector for each small region of the input face image and the registered face image. good.

The operation of the image recognition device 2 configured as above will be described with reference to FIG. It is assumed that the storage device 1004 stores a standard face image and a registered face image in advance. It is also assumed that the standard face image is preliminarily divided into small areas. It is also assumed that the registered face image is also divided into small areas in advance. Also, it is assumed that the registered face image is stored together with information representing the shielding pattern.

In FIG. 9, the dividing unit 23 first normalizes the input face image (input face image) (step S21).

Next, the dividing unit 23 divides the input face image normalized in step S21 into small regions (step S22).

Next, the reliability calculation unit 24 compares the small areas of the face image divided in step S22 with the small areas of the standard face image. Thereby, the reliability calculation unit 24 calculates the reliability of each small area (step S23).

Next, the determination unit 21 determines the masking pattern of the input face image based on whether or not the reliability distribution satisfies the masking condition for the group of small regions forming the middle region (step S24).

Specifically, as described above, if the reliability distribution of a middle region representing the upper half or the lower half of the face satisfies the masking condition, the determining unit 21 estimates that the middle region is the masking region. On the other hand, if the reliability distribution of the middle region representing the upper half or the lower half of the face does not satisfy the masking condition, the determination unit 21 estimates that the middle region is not the masking region. Then, when it is estimated that such a middle area is a shielding area, the determination unit 21 may determine a shielding pattern corresponding to the set middle area. For example, when the determination unit 21 estimates that the upper half area is a shielded area, it determines that the pattern is shielded by sunglasses. Further, when the determination unit 21 estimates that the lower half area is a shielded area, the determination unit 21 determines that it is a masked pattern.

Next, the identification unit 22 compares the corresponding small areas between the small areas obtained by dividing the input face image and the small areas forming the registered face image. Thereby, the identification unit 22 calculates the identification score between each small area (step S25).

As described above, the identification unit 22 may calculate the identification score from the distance or correlation between feature amount vectors or low-dimensional feature amount vectors.

Next, the identification unit 22 excludes the area based on the shielding pattern determined in step S24, and calculates an integrated score by integrating the identification scores between the small areas calculated in step S25 (step S26).

Here, if one or both of the corresponding small regions between the input face image and the registered face image are included in the exclusion region based on the shielding pattern of each image, the identification unit 22 calculates the identification score between the small regions. Not used for integration. Then, if both of the corresponding small regions are not included in the exclusion regions based on the shielding pattern of each image, the identification unit 22 uses the identification scores between the small regions for integration.

For example, the identification unit 22 may use the average value of the identification scores of the corresponding small regions as the integrated score. If no shielding pattern is determined for both the input face image and the registered face image, the identifying section 22 takes the average value of the identification scores of all the small regions as the integrated score. For example, when a mask shielding pattern is determined in one or both of the input face image and the registered face image, the identification unit 22 sets the lower half of the face as an exclusion area, and calculates the identification score of the small area in the upper half of the face. An average value may be used as an integrated score. Further, for example, the identification unit 22 may multiply the identification score of each target small region by the reliability of each small region, take an average, and use the weighted average value as the integrated score.

Thus, the image recognition device 2 ends its operation.

Next, effects of the second embodiment of the present invention will be described.

The image recognition apparatus according to the second embodiment of the present invention can further improve the recognition accuracy of face images including masked areas without increasing the cost of system construction and the load of identification processing.

The reason for this is that the dividing unit divides the face image into small regions, and the reliability calculation unit calculates the reliability of each small region by comparing the corresponding small regions between the face image and the standard face image. Because it does. This is because the determination unit estimates the masked area of the face image based on the distribution of the reliability for each small area, and determines the masked pattern based on the distribution of the masked area. This is because the identifying section excludes the area determined based on the masking pattern in the face image, identifies each small area, and identifies the entire face image based on each identification result.

As described above, in the present embodiment, rather than determining whether or not each small region constituting a face image is an occluded region based only on the reliability thereof, it is determined based on a wider range of reliability distributions. , to estimate the occluded region. As a result, the present embodiment can reduce the influence of errors in the shielding determination for each small area, and can estimate the shielding area more accurately. In this embodiment, since the shielding pattern is determined based on the estimated distribution of the shielding regions, the influence of errors in the shielding determination can be further reduced, and the shielding pattern can be determined with higher accuracy. Then, in the present embodiment, identification is performed by excluding an area determined according to the determined shielding pattern, instead of excluding the area estimated as the shielding area as it is. Therefore, this embodiment is more robust against occlusion determination errors. Then, in the present embodiment, such areas are excluded, and the face image is identified based on the identification result for each small area. As a result, according to the present embodiment, it is possible to realize shielding detection that is robust against the detection error of the shielding area, and to improve the recognition accuracy of the face image.

Furthermore, in the present embodiment, since the masked area and masking pattern are determined by comparison with a single standard face, a large amount of learning data, which is necessary for the determination of the masking area and masking pattern in the existing technology, can be obtained. does not require As a result, this embodiment can significantly reduce the cost of system construction.

In the present embodiment, an example has been mainly described in which the registered image is pre-stored in the storage device together with its shielding pattern and divided small area information. Not limited to this, in the present embodiment, when identifying an input face image, division of a registered image and determination of a shielding pattern may also be performed.

(Third Embodiment)
Next, a third embodiment of the present invention will be described in detail with reference to the drawings. In each drawing referred to in the description of this embodiment, the same reference numerals are assigned to the same configurations and steps that operate in the same manner as in the third embodiment of the present invention, and detailed descriptions of the present embodiment are given. Description is omitted.

First, FIG. 10 shows the configuration of an image recognition device 3 as a third embodiment of the present invention. In FIG. 10, the image recognition device 3 has an overall identifying section 35 and a switching section 36 in addition to the same configuration as the image recognition device 2 as the second embodiment of the present invention.

Here, the image recognition device 3 can be composed of hardware elements similar to those of the image recognition device 1 according to the first embodiment of the present invention described with reference to FIG. In this case, the overall identification unit 35 is composed of an output device 1006 and a CPU 1001 that loads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. The switching unit 36 is configured by the CPU 1001 that loads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. The hardware configuration of the image recognition device 3 and its functional blocks is not limited to the configuration described above.

Next, each functional block of the image recognition device 3 will be described.

The overall identification unit 35 identifies the face image using information representing the entire face in the face image. Specifically, for example, the overall identification unit 35 performs feature extraction from the entire face to obtain a feature amount vector. The overall identification unit 35 also obtains a feature amount vector for the entire face from the registered face image. Then, the overall identification unit 35 may calculate the overall identification score based on the feature amount vector of the entire face of the input face image and the feature amount vector of the entire face of the registered face image. In this case, the overall identification unit 35 outputs the overall identification score as the identification result.

The switching unit 36 switches between using the identifying unit 22 and the overall identifying unit 35 for identifying the input face image based on the determination result of the shielding pattern by the determining unit 21 .

The operation of the image recognition device 3 configured as described above will be described with reference to FIG. In the following, it is assumed that registered face images are stored in storage device 1004 along with information representing shielding patterns and small regions.

In FIG. 11, first, the image recognition device 3 operates in steps S21 to S24 in the same manner as the image recognition device 2 according to the second embodiment of the present invention to determine the shielding pattern of the input face image.

Next, the switching unit 36 switches subsequent processes based on the determined shielding pattern. Specifically, the switching unit 36 determines whether or not it is determined that at least one of the input face image and the registered face image has a shielding pattern regardless of the type (step S31).

Here, if it is determined that there is a shielding pattern in at least one of them, the switching unit 36 controls the identifying unit 22 to execute steps S25 and S26 in the same manner as in the second embodiment of the present invention. .

On the other hand, when it is determined that there is no shielding pattern in both the input face image and the registered face image, the switching section 36 controls the overall identification section 35 to execute the following step S32.

Here, the overall identification unit 35 performs identification using information on the entire face of the input face image and information on the entire face of the registered face image. Specifically, the overall identification unit 35 calculates the overall identification score using the feature amount vector obtained from the entire face of the input face image and the feature amount vector obtained from the entire face of the registered face image (step S32).

For example, the overall identification unit 35 may obtain a feature amount vector from the entire face of each image by feature extraction processing such as Gabor filter or LBP. Then, the overall identification unit 35 may calculate the overall identification score from the feature vector of the entire face of each of the input face image and the registered face image and the distance or correlation between them.

Thus, the image recognition device 3 ends its operation.

Next, effects of the third embodiment of the present invention will be described.

The image recognition apparatus according to the third embodiment of the present invention has the same effects as the second embodiment of the present invention, and furthermore, has the following features: can be

The reason for this is that when there is no masking pattern in either the input face image or the registered face image, the switching unit does not use the identification result for each small region, but uses the information for the entire face to perform processing. This is because the This is also because the switching unit switches the processing to perform identification using the identification result for each small region when at least one of the input face image and the registered face image has a masking pattern.

Here, when there is no shielding pattern and there is no need to exclude regions, there is a tendency that the identification result using the information of the entire face is more accurate than the integration of the identification results for each small region. Therefore, the present embodiment maintains high authentication accuracy when at least one of both images has a shielding pattern, and does not reduce matching accuracy even when neither of the images has a shielding pattern.

In the present embodiment, an example in which the identification unit and the overall identification unit calculate the identification score based on the distance and correlation between the feature amount vectors has been described. you can go

(Fourth embodiment)
Next, a fourth embodiment of the present invention will be described in detail with reference to the drawings. In each drawing referred to in the description of this embodiment, the same reference numerals are assigned to the same configurations and steps that operate in the same manner as in the third embodiment of the present invention, and detailed descriptions of the present embodiment are given. Description is omitted.

First, FIG. 12 shows the configuration of an image recognition device 4 as a fourth embodiment of the present invention. In FIG. 12, an image recognition device 4 replaces the identification unit 22 with an identification unit 42 and replaces the reliability calculation unit 24 with a reliability degree It differs in that it includes a calculation unit 44 , an overall identification unit 45 instead of the overall identification unit 35 , and further includes a feature extraction unit 47 .

Here, the image recognition device 4 can be configured by hardware elements similar to those of the image recognition device 1 according to the first embodiment of the present invention described with reference to FIG. In this case, the feature extracting unit 47 is configured by the CPU 1001 that reads computer programs and various data stored in the ROM 1003 and storage device 1004 into the RAM 1002 and executes them. The hardware configuration of the image recognition device 4 and its functional blocks is not limited to the configuration described above.

The feature extraction unit 47 extracts a feature amount vector for each small area of the face image divided by the division unit 23 . The various filters and techniques described above may be used to extract the feature vector. For example, the feature extraction unit 47 may extract Gabor features using a Gabor filter. Specifically, the feature extraction unit 47 calculates a feature amount vector for each small region normalized and divided by the division unit 23 for the input face image, and stores the vector in the storage device 1004 . Further, the feature extracting section 47 may also calculate feature amount vectors for each small area in advance for the standard face image and the registered face image. In this embodiment, the standard face image may be stored in the storage device 1004 as a feature amount vector for each small area, and the data of the face image itself may not be stored. Similarly, the registered face image may be stored in the storage device 1004 as a feature amount vector for each small area, and the data of the registered face image itself may not be stored.

The reliability calculation unit 44 is configured substantially in the same manner as the reliability calculation unit 24 in the second and third embodiments of the present invention. That is, the reliability calculation unit 44 calculates the reliability between the input face image and the standard face image using the feature vector for each corresponding small region. However, the reliability calculation unit 44 differs from the reliability calculation unit 24 in that the feature vector of each small region of the input face image and the standard face image is obtained from the storage device 1004 instead of being calculated.

The identifying section 42 is configured substantially in the same manner as the identifying section 22 in the second and third embodiments of the present invention. That is, the identification unit 42 calculates the identification score using the feature vector for each corresponding small area between the input face image and the registered face image. However, the identification unit 42 differs from the identification unit 22 in that the feature amount vector of each small region of the input face image and the registered face image is obtained from the storage device 1004 instead of being calculated. Further, the identification unit 42 may learn in advance a transformation matrix for transforming the feature amount vector to a lower dimension, as in the second to third embodiments of the present invention. In that case, the identification unit 42 calculates an identification score using a feature amount vector for each small area that has been converted to a lower dimension. For example, the identification unit 42 may calculate the normalized correlation value between the low-dimensional feature amount vectors of the corresponding small regions of the input face image and the registered face image as the identification score. Then, the identification unit 42 excludes any area based on the shielding pattern of the input face image and the registered face image, and integrates the identification results for each small area to calculate an integrated score.

The overall identification unit 45 calculates an overall identification score based on the connected feature amount vector of the input face image and the connected feature amount vector of the registered face image. Here, the concatenated feature quantity vector is generated by concatenating the feature quantity vectors of the respective small regions. For example, it is assumed that feature amount vectors of n1 rows and 1 column, n2 rows and 1 column, and n3 rows and 1 column are extracted for three small regions. n1, n2 and n3 are each positive integers. These may all have the same value, or at least some of them may have different values. In this case, the overall identification unit 45 can connect each feature amount vector to generate a (n1+n2+n3)-row, 1-column connected feature amount vector. In other words, the number of dimensions of the concatenated feature quantity vector is greater than that of the feature quantity vector of each small region. The order of concatenation is arbitrary.

In addition, the overall identification unit 45 also learns in advance a conversion matrix for converting the connected feature amount vector to a low-dimensional connected feature amount vector by linear discriminant analysis on the learning data for the connected feature amount vector. good too. In this case, the overall identification unit 45 may calculate the normalized correlation value between the low-dimensional connected feature amount vectors as the overall identification score.

The operation of the image recognition device 4 configured as described above will be described with reference to FIG. In the following description, it is assumed that the storage device 1004 stores a feature amount vector for each small area of the standard face image. Also, it is assumed that the feature amount vector for each small area of the registered face image is stored in the storage device 1004 together with the shielding pattern determined by the determination unit 21 in advance.

In FIG. 13, the dividing unit 23 first executes steps S21 and S22 in the same manner as in the second embodiment of the present invention to normalize and divide the input face image.

Next, the feature extraction unit 47 extracts a feature amount for each small area of the input face image and stores it in the storage device 1004 (step S41).

Next, the reliability calculation unit 44 calculates the feature vector of each small region of the input face image stored in the storage device 1004 in step S41 and the feature vector of each small region of the standard face image recorded in the storage device 1004. Quantity vectors are used to calculate the reliability between corresponding small regions (step S42).

Next, the determination unit 21 executes step S24 to determine the shielding pattern as in the second embodiment of the present invention.

Next, when the switching unit 36 determines that there is a shielding pattern in one or both of the input face image and the registered face image, the identification unit 42 executes steps S43 and S26.

Here, for the input face image and the registered face image, the identification unit 42 uses the feature vectors of the small regions stored in the storage device 1004 to calculate the identification score between the corresponding small regions (step S43). .

For example, as described above, the identification unit 42 may calculate a normalized correlation value as an identification score between low-dimensional feature amount vectors obtained by converting feature amount vectors of corresponding small regions into low-dimensional ones. Then, the identification unit 42 executes step S26 to calculate an integrated score, as in the second embodiment of the present invention.

On the other hand, when the switching unit 36 determines that there is no shielding pattern in both the input face image and the registered face image, the overall identification unit 45 executes steps S44 and S45.

Here, for the input face image and the registered face image, the overall identification unit 45 generates a concatenated feature quantity vector by concatenating the feature quantity vectors of the small regions stored in the storage device 1004 (step S44).

Then, for the input face image and the registered face image, the overall identification unit 45 calculates the overall identification score by comparing the connected feature amount vectors (step S45).

For example, as described above, the overall identification unit 45 may convert the concatenated feature amount vectors to lower dimensions, and then calculate the normalized correlation value between the low-dimensional concatenated feature amount vectors as the overall identification score.

Thus, the image recognition device 4 ends its operation.

Next, effects of the fourth embodiment of the present invention will be described.

The image recognition apparatus according to the fourth embodiment of the present invention achieves higher processing efficiency and speed while maintaining high recognition accuracy for face images including masked areas.

The reason for this is that the feature extraction unit extracts and stores feature amount vectors for each of the small regions that form the input face image, the standard face image, and the registered face image. As a result, the reliability calculation unit, the identification unit, and the overall identification unit can commonly use the already-extracted feature amount vectors, and each process can be made more efficient and faster.

In each of the second to fourth embodiments of the present invention described above, examples have been mainly described in which the identification unit and the overall identification unit identify an input face image by matching with a registered image. Without being limited to this, the identifying unit and overall identifying unit of each embodiment may perform gender estimation, posture estimation, facial expression recognition, etc. of a person represented by an input face image without using a registered image. In this way, each embodiment can also be applied to various identification processes that do not use a registered image for a face image that includes a masked area.

Further, in each of the second to fourth embodiments of the present invention described above, the determining unit, the identifying unit, and the overall identifying unit use the distance or correlation between the feature amount vectors to determine the shielding pattern and input face information. The description has focused on an example of image identification. Without being limited to this, these functional blocks of each embodiment may calculate reliability or identification score by comparing regions by other methods. In this case, the feature extraction unit according to the fourth embodiment of the present invention may calculate and store information used for comparison with respect to the corresponding small area of the image.

Further, in each embodiment of the present invention described above, shielding by a mask or sunglasses has been described as an example of a shielding pattern, but the type of shielding pattern is not limited.

In each embodiment of the present invention described above, the input face image, the standard face image, and the registered face image refer to regions representing faces included in the corresponding images. These face images may be obtained by cutting out the face area in advance, or by cutting out during processing.

Further, in each of the above-described embodiments of the present invention, an example in which each functional block of the image recognition apparatus is implemented by a CPU executing a computer program stored in a storage device or ROM has been mainly described. Not limited to this, some or all of each functional block, or a combination thereof may be implemented by dedicated hardware.

Moreover, in each embodiment of the present invention described above, the functional blocks of the image recognition device may be implemented by being distributed to a plurality of devices.

Further, in each embodiment of the present invention described above, the operation of the image recognition apparatus described with reference to each flowchart may be stored in a computer storage device (storage medium) as the image recognition program of the present invention. good. Then, the computer program may be read and executed by the CPU. In such a case, the present invention is constituted by such computer program code or storage medium.

Moreover, each embodiment mentioned above can be combined suitably and implemented.

Moreover, the present invention is not limited to the above-described embodiments, and can be implemented in various modes.

In addition, part or all of each of the embodiments described above can also be described in the following additional remarks, but is not limited to the following.
(Appendix 1)
a determination unit that determines a masking pattern of the face in the face image by comparing an image representing the face (face image) with a standard image representing the face (standard face image);
an identification unit that identifies the face image by excluding a region based on the shielding pattern in the face image;
An image recognition device with
(Appendix 2)
The determining unit selects the shielding pattern based on a reliability distribution based on a comparison between each small region forming the face image and a small region corresponding to each small region of the face image in the standard face image. 1. The image recognition device according to appendix 1, characterized in that:
(Appendix 3)
The determination unit is configured to, for a middle area obtained by putting together the small area group within a predetermined range, based on the middle area when a reliability distribution of the small area group constituting the middle area satisfies a predetermined shielding condition. 2. The image recognition device according to appendix 2, wherein the shielding pattern is determined.
(Appendix 4)
The determination unit defines an upper face region, which is a group of small regions above the nose in the face image, as the middle region, and determines that if the reliability distribution in the middle region is the shielding condition, the sunglasses are used. 3. The image recognition device according to appendix 3, wherein the image recognition device determines that the pattern is a shielding pattern.
(Appendix 5)
The determination unit defines a face lower region, which is a group of small regions located below the nose in the face image, as the middle region, and if the reliability distribution satisfies the shielding condition in the middle region, the mask is determined. The image recognition device according to appendix 3 or appendix 4, wherein the image recognition apparatus determines that the pattern is a shielding pattern by .
(Appendix 6)
a dividing unit that divides the face image into the small areas;
a reliability calculation unit that calculates the reliability by comparing the corresponding small areas of the face image and the standard face image;
The image recognition device according to any one of appendices 2 to 5, further comprising:
(Appendix 7)
The identification unit identifies each of the small regions other than the region based on the shielding pattern in the face image, and identifies the face image based on the identification result of each of the small regions. The image recognition device according to any one of appendices 2 to 6.
(Appendix 8)
an overall identification unit that identifies the face image using information of the entire face;
a switching unit that switches between the identification unit and the overall identification unit to be used based on the determination result of the determination unit;
The image recognition device according to any one of appendices 1 to 7, further comprising:
(Appendix 9)
9. The image recognition apparatus according to any one of appendices 1 to 8, wherein a face image representing an average face obtained by averaging faces of a plurality of persons is used as the standard face image.
(Appendix 10)
9. The image recognition apparatus according to any one of appendices 1 to 8, characterized in that, as the standard face image, a face image satisfying a predetermined condition in degree of similarity to face images of a plurality of persons is used.
(Appendix 11)
9. The image recognition apparatus according to any one of appendices 1 to 8, wherein a face image representing a standard facial artifact is used as the standard face image.
(Appendix 12)
12. The image recognition apparatus according to any one of appendices 1 to 11, wherein information represented by a feature amount vector is used as the face image and the standard face image.
(Appendix 13)
determining a masking pattern of the face in the face image by comparing an image representing the face (face image) with an image representing a standard face (standard face image);
An image recognition method for identifying the face image by excluding a region based on the shielding pattern in the face image.
(Appendix 14)
a determination step of determining a masking pattern of the face in the face image by comparing an image representing the face (face image) with a standard image representing the face (standard face image);
an identification step of identifying the face image by excluding a region based on the shielding pattern in the face image;
An image recognition program that causes a computer device to execute

1, 2, 3, 4 image recognition device 11, 21 determination unit 12, 22, 42 identification unit 23 division unit 24, 44 reliability calculation unit 35, 45 overall identification unit 36 switching unit 47 feature extraction unit 1001 CPU
1002 RAMs
1003 ROMs
1004 storage device 1005 imaging device 1006 output device

Claims (11)

a determination unit that determines a shielded region that is shielded in an input image that is an image representing a face;
an identification unit that identifies the input image using a region other than an exclusion region corresponding to the blocking pattern, which is pre- associated with the blocking pattern based on the blocking region;
Information processing device with The determination unit determines a reliability distribution based on a comparison between a plurality of small regions forming the input image and small regions corresponding to each of the small regions of the input image in a standard image representing a standard face. The information processing apparatus according to claim 1, wherein the shielding pattern is determined based on. The determination unit determines the shielding pattern when a reliability distribution of the small areas constituting the medium area satisfies a predetermined shielding condition for a medium area that is a small area group obtained by collecting the small areas of a predetermined range. 3. The information processing apparatus according to claim 2, wherein the determination is made. The determination unit defines an upper face area obtained by collecting the small areas above the nose in the input image as the middle area, and if the reliability distribution satisfies the shielding condition in the middle area, the shielding by sunglasses is performed. 4. The information processing apparatus according to claim 3, wherein the pattern is determined. The determination unit defines a lower face region obtained by collecting the small regions below the nose in the input image as the middle region, and uses a mask when the reliability distribution satisfies the shielding condition in the middle region. 5. The information processing apparatus according to claim 3, wherein the pattern is determined to be a shielding pattern. a dividing unit that divides the input image into the small regions;
a reliability calculation unit that calculates the reliability by comparing the corresponding small areas of the input image and the standard image;
The information processing apparatus according to any one of claims 2 to 5, further comprising: The identification unit identifies the input image for each of the small regions other than the region determined based on the shielding pattern in the input image, and identifies the input image based on the identification result for each of the small regions. 7. The information processing apparatus according to any one of claims 2 to 6. an overall identification unit that identifies the input image using information on the entire face;
a switching unit that switches between the identification unit and the overall identification unit to be used based on the determination result of the determination unit;
The information processing apparatus according to any one of claims 1 to 7, further comprising: The identification unit identifies the input image by excluding from the input image the exclusion area associated with the shielding pattern based on the shielding area,
The information processing apparatus according to any one of claims 1 to 8, wherein the exclusion area is an area different from the shielded area determined by the determination unit. Determining an occluded area that is occluded in an input image that is an image representing a face,
An information processing method for identifying the input image using areas other than exclusion areas corresponding to the shielding pattern, which are pre- associated with the shielding pattern based on the shielding area. A process of determining a shielded region that is shielded in an input image that is an image representing a face;
a process of identifying the input image using a region other than an exclusion region according to the blocking pattern, which is pre- associated with the blocking pattern based on the blocking region;
An information processing program that causes a computer device to execute

Images