| 摘要 |
The present invention relates to a virtual viewpoint depth image synthesis method for generation of a computer generated hologram (CGH) capable of obtaining a virtual viewpoint depth image from depth images of a plurality of reference view cameras. The method comprises the steps of: (a) acquiring depth images (hereinafter, referred to as a reference view depth image) from at least two depth cameras (hereinafter, referred to as reference view cameras); (b) extracting parameters of the reference view cameras through camera calibration; (c) determining a camera, closest to a virtual viewpoint camera, (hereinafter, referred as a primary reference view camera) and a closest camera among the reference view cameras, positioned in a direction opposite to a position of the primary reference view camera, (hereinafter, referred as a secondary reference view camera); (d) by using a parameter of the primary reference view camera, generating a virtual viewpoint depth image from the reference view depth image by the primary reference view camera (hereinafter, referred as a primary reference view depth image); (e) compensating an occlusion region of the virtual viewpoint depth image by means of a secondary reference view depth image; and (f) generating a hologram from the compensated virtual viewpoint depth image. By the above-described virtual viewpoint depth image synthesis method, the virtual viewpoint depth image is generated from a primary reference view image closest to a virtual viewpoint, and the occlusion region is compensated by a reference image opposite to the primary reference view image. As a result, a virtual viewpoint depth image with high reliability is generated and image quality is improved, thereby generating a more accurate and superior virtual viewpoint depth image. [Reference numerals] (AA) Start; (BB) End; (S10) Acquire reference view depth images; (S20) Extract parameters of reference view cameras; (S30) Determine a primary reference view camera and a secondary reference view camera; (S40) Generate a virtual view depth image from a primary reference view depth image; (S50) Compensate an occlusion region of the virtual view depth image by means of a secondary reference view depth image; (S60) Apply a hole filling technique to the compensated depth image to generate a final virtual view depth image; (S70) Generate a hologram from the final virtual view depth image |
