作者: Walid Barhoumi , Mohammed Chafik Bakkay , Ezzeddine Zagrouba
DOI: 10.1007/S11760-011-0273-1
关键词:
摘要: We propose in this work to synthesize online the visual content of a complex video into multiple sprites, while detecting on-the-fly each sprite’s limits. For received frame, physically meaningful camera rotation angles and focal lengths are firstly estimated from frame-to-sprite homography. These physical parameters then used by thresholding-based assignment stage confirm whether frame can be added without distortions on one sprites that have been built up instant. In case, selected sprite will updated accordingly applying direct warping advanced median blending only newly discovered background content. Else, new must defined selecting current as reference frame. Extensive experiments accomplished challenging videos demonstrate proposed approach process long preventing accumulation alignment errors.