Image generating apparatus, image generating method and image generating program

Abstract

An image generating apparatus for generating a visual field image seen from a photographed person who is photographed in at least one part of plural two-dimensional images, based on the plural two-dimensional images captured by plural image capturing apparatuses, comprises a photographed person's position detector for detecting a position of the photographed person, based on at least two of the plural two-dimensional images; a photographed person's direction detector for detecting a face direction of the photographed person; a three-dimensional information generating section for generating three-dimensional information of space that is photographed, based on at least two of the two-dimensional images; and an image generating section for generating the visual field image seen from the position of the photographed person along the face direction, based on the three-dimensional information, the position of the photographed person, and the face direction of the photographed person.

1 . An image generating apparatus for generating an image including a photographed person in at least one part of the plural two-dimensional images, based on plural two-dimensional images captured by plural image capturing apparatuses, comprising:
 a photographed person's position detector for detecting a position of the photographed person;   a photographed person's direction detector for detecting a face direction of the photographed person;   a three-dimensional information generating section for generating three-dimensional information of space that is photographed, based on at least two of the two-dimensional images;   a viewpoint deciding section for deciding a viewpoint which is substantially on an extension along the face direction of the photographed person, based on the position and face direction of the photographed person; and   an image generating section for generating an image including the photographed person, seen from the viewpoint, based on the three-dimensional information and the position of the viewpoint.   
   
   
       2 . The image generating apparatus as claimed in  claim 1 , wherein said photographed person's position detector detects the position of the photographed person based on at least two images of the two-dimensional images,
 said photographed direction detector detects the visual direction of the photographed person as the face direction, and   said viewpoint deciding section decides the viewpoint which is substantially on an extension along the visual direction of the photographed person, based on the position and visual direction of the photographed person.   
   
   
       3 . The image generating apparatus as claimed in  claim 2 , wherein said photographed person's direction detector detects the visual direction of the photographed person, based on at least one of the plural two-dimensional images. 
   
   
       4 . The image generating apparatus as claimed in  claim 1 , wherein said photographed person's position detector detects each of the positions of the plural photographed persons and calculates an average position by averaging the positions of the plural photographed persons,
 said photographed person's direction detector detects each of the face directions of the plural photographed persons and calculates an average direction by averaging the face directions of the plural photographed persons,   said viewpoint deciding section decides the viewpoint which is substantially on an extension, extending from the average position along the average direction, and   said image generating section generates an image including the plural photographed persons seen from the position of the viewpoint, based on the three-dimensional information and the position of the viewpoint.   
   
   
       5 . The image generating apparatus as claimed in  claim 1 , further comprising:
 a central region detector for detecting a central region so that a density regarding the number of extensions which extend from each of the plural photographed persons along each of the face directions of the photographed persons, is equal to or greater than predetermined density; and   a main subject person selector for selecting photographed persons from the plural photographed persons as main subject persons so that the extensions which extend from the photographed person along the face directions of the photographed persons intersect at the central region; wherein   said photographed person's position detector calculates an average position by averaging the positions of the main subject persons based on at least two of the plural two-dimensional images,   said photographed person's direction detector calculates an average direction by averaging the face directions of the main subject persons, and   said image generating section generates the visual field image seen from the average position along the average direction, based on the three-dimensional information, the average position, and the average direction.   
   
   
       6 . The image generating apparatus as claimed in  claim 1 , wherein said photographed person's position detector detects each of the positions of the plural photographed persons, and calculates an average position by averaging the positions of the plural photographed persons, further comprising:
 a main subject person selector for selecting a photographed person as a main subject person from the plural photographed persons so that the selected position of the photographed person is nearest to the average position,   wherein said photographed person's direction detector detects the face direction of the main subject person,   said viewpoint deciding section decides a viewpoint which is substantially be on an extension extending from the position of the main subject person along the face direction of the main subject person, based on the position and face direction of the main subject person, and   said image generating section generates an image including the plural photographed persons seen from the position of the viewpoint, based on the three-dimensional information and the position of the viewpoint.   
   
   
       7 . The image generating apparatus as claimed in  claim 1 , wherein said image generating section generates an images including both the photographed person seen from the position of the viewpoint and a predetermined subject, based on the three-dimensional information and the position of the viewpoint. 
   
   
       8 . An image generating method for generating a visual field image by a computer, which the image is seen from a photographed person who is photographed in at least one part of the plural two-dimensional images, based on plural two-dimensional images captured by plural image capturing apparatuses, comprising:
 a photographed person's position detecting step for detecting a position of the photographed person, by the computer;   a photographed person's direction detecting step for detecting a face direction of the photographed person, by the computer;   a three-dimensional information generating step for generating section for generating three-dimensional information of space that is photographed, based on at least two of the two-dimensional images, by the computer;   a viewpoint deciding step for deciding a viewpoint which is substantially on an extension of the face direction along the photographed person, based on the position and direction of the photographed person, by the computer; and   an image generating step for generating an image including the photographed person seen from the viewpoint, based on the three-dimensional information and the position of the viewpoint, by the computer.   
   
   
       9 . A computer readable medium storing thereon a program for generating an image including a photographed person in at least one part of the plural two-dimensional images, based on plural two-dimensional images captured by plural image capturing apparatuses, the program comprising:
 a photographed person's position detecting module for detecting a position of the photographed person;   a photographed person's direction detecting module for detecting a face direction of the photographed person;   a three-dimensional information generating module for generating three-dimensional information of space that is photographed, based on at least two of the two-dimensional images;   a viewpoint deciding module for deciding a viewpoint which is substantially on an extension along the face direction of the photographed person, based on the position and face direction of the photographed person; and   an image generating module for generating an image including the photographed person, seen from the viewpoint, based on the three-dimensional information and the position of the viewpoint.