Image processing apparatus and method for generating high quality image
An image processing apparatus includes: an acquisition unit configured to acquire a plurality of images each capturing an identical target and having a different attribute; a derivation unit configured to derive features from the plurality of images, using a first neural network; an integration unit configured to integrate the features derived from the plurality of images; and a generation unit configured to generate a higher quality image than the plurality of images from the feature integrated by the integration unit, using a second neural network.
1. An image processing apparatus comprising:
one or more processors; and
a memory coupled to the one or more processors, the memory having stored thereon instructions which, when executed by the one or more processors, cause the image processing apparatus to:
acquire a plurality of images being obtained by capturing an identical target from different viewpoint positions by a plurality of image capturing devices;
derive features from each of the plurality of images using corresponding one of first neural networks each provided for respective one of the viewpoint positions of the images;
integrate the features derived from the plurality of images; and
generate a high quality image of which quality is higher than that of each of the plurality of images, from the integrated features, using a second neural network.
2. The image processing apparatus according to claim 1 , wherein integrating the features includes determining a connectivity relationship of individual neural networks on a basis of a geometric positional relationship between viewpoint positions of the plurality of images and a viewpoint position of the high quality image.
3. The image processing apparatus according to claim 1 , wherein the one or more processors cause the image processing apparatus to integrate the features by determining a connectivity relationship of individual neural networks on a basis of case images for learning.
4. The image processing apparatus according to claim 1 , wherein the one or more processors cause the image processing apparatus to integrate the features in a plurality of stages, and determines an order of integration of the features on a basis of proximity of the viewpoint positions of the images corresponding to the features.
5. An image processing method comprising:
acquiring a plurality of images being obtained by capturing an identical target from different viewpoint positions by a plurality of image capturing devices;
deriving features from each of the plurality of images using corresponding one of first neural networks each provided for respective one of the viewpoint positions of the images;
integrating the features derived from the plurality of images; and
generating a high quality image of which quality is higher than that of each of the plurality of images, from the integrated features, using a second neural network.
6. A non-transitory computer-readable storage medium storing a program causing a computer to execute a method of image processing, the method comprising:
acquiring a plurality of images being obtained by capturing an identical target from different viewpoint positions by a plurality of image capturing devices;
deriving features from each of the plurality of images using corresponding one of first neural networks each provided for respective one of the viewpoint positions of the images;
integrating the features derived from the plurality of images; and
generating a high quality image of which quality is higher than that of each of the plurality of images, from the integrated features, using a second neural network.