Synthesis of transformed image views

A system for view synthesis comprising:

• an image transformation circuit to transform a reference image frame corresponding to a first viewpoint, to a transformed image frame corresponding to a second viewpoint, the reference image frame and the transformed image frame each comprising a plurality of pixels associated with color values and depth values;
• an inverse warping circuit to calculate a mapping from the pixels of the transformed image frame back to corresponding pixels of the reference image frame;
• a hole detection circuit to detect holes in the transformed image frame based on discontinuities between the depth values of the pixels of the transformed image frame and the depth values of the corresponding pixels mapped back to the reference image frame;
• and
• a hole filling circuit to in-fill one of the detected holes, using a sampling of selected neighboring pixels from the reference image frame, to synthesize a view based on the transformed image frame.

At least one non-transitory computer readable storage medium having instructions encoded thereon that, when executed by one or more processors, result in the following operations for view synthesis, the operations comprising:

• transforming a reference image frame corresponding to a first viewpoint, to a transformed image frame corresponding to a second viewpoint, the reference image frame and the transformed image frame each comprising a plurality of pixels associated with color values and depth values;
• calculating a mapping from the pixels of the transformed image frame back to corresponding pixels of the reference image frame;
• detecting holes in the transformed image frame based on discontinuities between the depth values of the pixels of the transformed image frame and the depth values of the corresponding pixels mapped back to the reference image frame;
• and
• performing in-filling of one of the detected holes, using a sampling of selected neighboring pixels from the reference image frame, to synthesize a view based on the transformed image frame.