The compression of color image and depth image for three-dimensional stereo image

• Main Authors: Shu-Fen Wu, 吳淑芬
• Other Authors: 范育成
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/yrks4z

碩士 === 國立臺北科技大學 === 電腦與通訊研究所 === 98 === Due to the rapid growth of 3D TV and 3D consuming electronic products, the issue of compressed stereo image for easier transmission and store catches much attention. ATTEST 3D TV System uses single color image and its depth image as three-dimensional stereo image whose data is much slighter than traditional 2D image files. Most video compression techniques target at time/space prediction and have high compression efficiency on color image. However, video compression techniques do not have much effect on depth image. This research discusses “The compression of color image and depth image for three-dimensional stereo image.” It deals with information of color image first. Motion vector would be known by variable block size motion estimation. Searching of variable block size would help to get object’s movement more precisely. However, wrong motion vector figure would appear if we use the same way, estimation of movement, on depth image. Therefore, we need to make some adjustment on the reference frame of depth image first. Then, adjusted depth map based motion estimated movement is performed to get true motion vector. Depth map parameter is adjusted during the procedure. Next, instead of 2D depth map search, 3D depth map search is done to achieve efficiently depth image compression. As a result, both color image and depth image would have fine compression effect by depth parameter adjustment.