Determination method, determination program, and information processing apparatus

1 . A determination method comprising:

acquiring a plurality of images which are captured at different times by a camera;

detecting a first image region of a face of a person in each of the plurality of images;

generating, from each of the plurality of images, a respective rectangular image by rotationally transforming the image so that a line segment coupling left and right eyes of the person represented in the first image region is parallel to one end of the rectangular image;

detecting a first linear image region and a second linear image region in each of the plurality of rectangular images, wherein the first linear image region is detected by a detection process starting from a pixel column parallel to the line segment, and the second linear image region is detected by a detection process starting from a pixel column perpendicular to the line segment; and

determining whether the face of the person is a display object based on a state of change in a positional relationship between the first image region and the first linear image region and a state of change in a positional relationship between the first image region and the second linear image region, between the plurality of images.

2 . The determination method according to claim 1 , further comprising:

detecting the first image region in each of the plurality of images;

generating, for the respective plurality of images, a plurality of edge intensity images in which an edge intensity at a position of each pixel is represented as a value of each pixel; and

detecting the first linear image region in each of the plurality of images by using the plurality of edge intensity images.

3 . The determination method according to claim 2 , further comprising:

selecting a pixel adjacent to a specific pixel and having a maximum edge intensity from an adjacent pixel column to a pixel column including the specific pixel with respect to a plurality of remaining edge intensity images obtained by excluding the first image region from each of the plurality of edge intensity images;

performing the selecting the pixel for each pixel column by starting from a pixel column adjacent to a first pixel column which is arranged at one end of the respective remaining edge intensity images while setting a pixel having a maximum edge intensity among pixels included in the first pixel column as a first specific pixel; and

detecting the first linear image region in each of the plurality of images by detecting an image region which is formed by a specific pixel group obtained by repeating the selecting the pixel in each of the plurality of remaining edge intensity images.

4 . The determination method according to claim 1 , wherein in the determining whether the face of the person is the display object, whether the change in the positional relationship is synchronized between the plurality of images is determined, and when determining that the change is not synchronized, the face of the person is determined not to be the display object.

5 . The determination method according to claim 4 , further comprising:

determining whether the change in the positional relationship is synchronized based on a similarity of a motion vector for each of the first image region and the second linear image region which are calculated using the plurality of images.

6 . The determination method according to claim 1 , wherein in the determining whether the face of the person is the display object, whether the change in the positional relationship is synchronized between the plurality of images is determined, and when determining that the change is synchronized, whether the first linear image region represents a contour of the display object, and when determining that the first linear image region represents the contour of the display object, the face of the person is determined to be the display object.

7 . The determination method according to claim 6 , wherein the determination as to whether the linear image region represents the contour of the display object is performed using at least one of an edge intensity at a position of each pixel included in the first linear image region, an inclination of a straight line which approximates the first linear image region with respect to a line segment which couples positions of left and right eyes of the person represented in the first image region, and a distance of the straight line from a position of a nose of the person represented in the first image region.

8 . The determination method according to claim 7 , wherein further comprising:

calculating a likelihood that the first linear image region represents the contour of the display object using the edge intensity at the position of each pixel included in the first linear image region, the inclination and the distance; and

determining whether the first linear image region represents the contour of the display object based on a result of magnitude comparison between the likelihood and a predetermined likelihood value.

9 . A non-transitory computer-readable recording medium storing a determination program for causing a computer to execute a process of:

acquiring a plurality of images which are captured at different times by a camera;

detecting a first image region of a face of a person in each of the plurality of images;

generating, from each of the plurality of images, a respective rectangular image by rotationally transforming the image so that a line segment coupling left and right eyes of the person represented in the first image region is parallel to one end of the rectangular image;

detecting a first linear image region and a second linear image region in each of the plurality of rectangular images, wherein the first linear image region is detected by a detection process starting from a pixel column parallel to the line segment, and the second linear image region is detected by a detection process starting from a pixel column perpendicular to the line segment; and

determining whether the face of the person is a display object based on a state of change in a positional relationship between the first image region and the first linear image region and a state of change in a positional relationship between the first image region and the second linear image region, between the plurality of images.

10 . An information processing apparatus comprising:

a memory; and

a processor coupled to the memory and configured to:

acquire a plurality of images which are captured at different times by a camera;

detecting a first image region of a face of a person in each of the plurality of images;

generating, from each of the plurality of images, a respective rectangular image by rotationally transforming the image so that a line segment coupling left and right eyes of the person represented in the first image region is parallel to one end of the rectangular image;

detecting a first linear image region and a second linear image region in each of the plurality of rectangular images, wherein the first linear image region is detected by a detection process starting from a pixel column parallel to the line segment, and the second linear image region is detected by a detection process starting from a pixel column perpendicular to the line segment; and

determine whether the face of the person is a display object based on a state of change in a positional relationship between the first image region and the first linear image region and a state of change in a positional relationship between the first image region and the second linear image region, between the plurality of images.