Multi-viewpoint image generation apparatus and method

摘要

A viewpoint position determination unit is configured to determine whether a viewpoint indicated by a viewpoint array signal is closer to a left-eye viewpoint or a right-eye viewpoint. A coefficient generation unit is configured to generate a coefficient based on the viewpoint array signal. A first selector is configured to select either a left-eye viewpoint image signal or a right-eye viewpoint image signal. A second selector is configured to select either a left-eye viewpoint depth value or a right-eye viewpoint depth value. A multiplier is configured to multiply the left-eye viewpoint depth value or the right-eye viewpoint depth value by the coefficient. A pixel shift unit is configured to carry out a pixel shift on the left-eye viewpoint image signal and the right-eye viewpoint image signal according to the depth value output from the multiplier, so as to generate a multi-viewpoint image signal.

主权项

1. A multi-viewpoint image generation method, comprising:
estimating concave-convex information of an object in an image based on a left-eye viewpoint image signal so as to generate a left-eye viewpoint concave-convex value; estimating the concave-convex information of the object in the image based on a right-eye viewpoint image signal so as to generate a right-eye viewpoint concave-convex value; adding a basic depth model that indicates a depth value of a predetermined scene structure in a frame to the left-eye viewpoint concave-convex value so as to generate a left-eye viewpoint depth value; adding the basic depth model to the right-eye viewpoint concave-convex value so as to generate a right-eye viewpoint depth value; generating a viewpoint array signal that maps one of a set of viewpoints to a pixel position in the frame; determining whether the viewpoint indicated by the viewpoint array signal is closer to a left-eye viewpoint or a right-eye viewpoint; generating a coefficient that is a positive number and increases in size as a distance of the viewpoint from a right side of the left-eye viewpoint or the right-eye viewpoint increases, and generating a coefficient that is a negative number and increases in size as a distance of the viewpoint from a left side of the left-eye viewpoint or the right-eye viewpoint increases; selecting only one of either the left-eye viewpoint image signal or the right-eye viewpoint image signal depending on whether the viewpoint indicated by the viewpoint array signal is closer to the left-eye viewpoint or the right-eye viewpoint, in both cases when generating a multi-viewpoint image signal including an image signal of a viewpoint outside the left-eye viewpoint or the right-eye viewpoint, and when generating a multi-viewpoint image signal including an image signal of a viewpoint between the left-eye viewpoint and the right-eye viewpoint; selecting only one of either the left-eye viewpoint depth value or the right-eye viewpoint depth value depending on whether the viewpoint indicated by the viewpoint array signal is closer to the left-eye viewpoint or the right-eye viewpoint, in both cases when generating a multi-viewpoint image signal including an image signal of a viewpoint outside the left-eye viewpoint or the right-eye viewpoint, and when generating a multi-viewpoint image signal including an image signal of a viewpoint between the left-eye viewpoint and the right-eye viewpoint; multiplying, by the coefficient, the left-eye viewpoint depth value or the right-eye viewpoint depth value selected; and carrying out a pixel shift on the left-eye viewpoint image signal and the right-eye viewpoint image signal according to the left-eye viewpoint depth value or the right-eye viewpoint depth value multiplied by the coefficient, so as to generate a multi-viewpoint image signal.