CPC H04N 19/597 (2014.11) [H04N 19/182 (2014.11)] | 9 Claims |
9. An evaluation device for evaluating coding quality of coded data of an image for a first viewpoint in a multi-viewpoint image, the evaluation device comprising:
a processor; and
a storage medium having computer program instructions stored thereon, when executed by the processor, perform to:
evaluate coding quality of coded data relating to the first viewpoint on a pixel-by-pixel basis using a difference value between a pixel value in an original image for the first viewpoint and a corresponding pixel value in the coded data for the first viewpoint, and a difference value between pixel value in the original image for a second viewpoint that is different from the first viewpoint and a corresponding pixel value in the coded data for the second viewpoint;
reflect evaluation relating to a third viewpoint that is different from the first viewpoint and the second viewpoint and is not included in images included in the multi-viewpoint image, in the evaluation of the coding quality of the coded data relating to the first viewpoint, by using the pixel value of the original image for the second viewpoint and the pixel value obtained from the coded data relating to the second viewpoint, wherein the third viewpoint is an arbitrary viewpoint included in an aggregate of viewpoints located between the first viewpoint and the second viewpoint;
evaluate coding quality of the coded data relating to the first viewpoint, using a difference value between a pixel value of an image for the third viewpoint based on the original image for the first viewpoint and the original image for the second viewpoint and a pixel value of an image for the third viewpoint based on the pixel value obtained from the coded data relating to the first viewpoint and the pixel value obtained from the coded data relating to the second viewpoint; and
evaluate coding quality of the coded data relating to the first viewpoint, using a value of SELBD calculated according to the following evaluation formula:
where W and H represent a number of pixels in a horizontal direction and a number of pixels in a vertical direction in the original image for the first viewpoint, respectively, and d(x, y) represents a difference value between a pixel value in the original image for the first viewpoint and a pixel value in a decoded image for the first viewpoint and d (x,y) represents a difference value between a pixel value in the original image for the second viewpoint and a pixel value in a decoded image for the second viewpoint.
|