Automatic video mosaicing using hierarchical compositing
Abstract — In this paper the problem of constructing mosaic from video sequences taken by rotating cameras is addressed. Video mosaicing find its applications in many popular arenas including video compression, virtual environments and panoramic photography. To construct a mosaiced image from a sequence of frames we need to obtain geometric transformation between successive frames. To establish this step we utilize a feature-based approach. Detected features are match through a block based method and then we estimate the mapping function parameters according to obtained feature matchings. A 2-pass stitching algorithm is then used to combine frames. Warped frames are combined in a hierarchical form of two or more levels to create submosaics. Then this submosaic images are mapped to the final mosaic surface and by employing a bilinear interpolation and an effective blending function a smooth and pleasing mosaic image is constructed. Experimental results are provided for several video sequences and the results obtained indicate the effectiveness of our method.
S.A Hoseini S. Jafari
school of electrical and computer engineering shiraz university Shiraz, Iran
国际会议
2010 International Conference on Signal and Information Processing(2010年IEEE信号与信息处理国际会议 ICSIP2010)
长沙
英文
466-469
2010-12-14(万方平台首次上网日期,不代表论文的发表时间)