Output details
11 - Computer Science and Informatics
Oxford Brookes University
Multiframe Motion Segmentation with Missing Data Using PowerFactorization and GPCA
<23> Identification of independently moving objects seen in a sequence of images is a hard problem that has been well studied. This problem is further complicated by missing data, where points cannot be successfully tracked through the whole
image sequence. This paper introduces a particularly simple and novel method for filling in missing data by a new algorithm known as PowerFactorization. Through simultaneous dimension reduction, the problem is made tractable
by simple linear techniques. Since the PowerFactorization method is applicable to different problems, this paper is well cited by papers in many different areas.