Font Size: a A A

Digital Video Repair A Number Of Key Technologies

Posted on:2007-12-03Degree:DoctorType:Dissertation
Country:ChinaCandidate:A J ChenFull Text:PDF
GTID:1118360185496363Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
Film is an invaluable artistic and cultural heritage in the long time of human history. However, many old films are being lost or aged because film is in a fragile state and on a volatile medium. It is extremely needed to conserve and restore these old films. On the other hand, artistic charm and commercial value of these old films have attracted more and more people when the restored films are showed to the public again. At the same time, these stored old films can provide abundant program source for digital TV and films.With the development of digital technology, digital conservation and restoration has become gradually the main method for conserving and restoring old films. Digital restoration is a technology using digital image processing to restore degradations existing in old films such as flicker, blotch, line-scratch, fade and so on. To develop a set of efficient and practical algorithms for correcting these artifacts is an attractive topic of research. Not too few research results of digital restoration have been reported in the overseas webs. Nevertheless no research work or report can be found in China. So the research on digital video restoration for old film is considered as the main content in this dissertation, where flicker removal and blotch removal are the emphases of research.The innovations in this dissertation can be listed as follows:1. Based on the smoothness property of flicker, a method detecting local motion is proposed. The flicker parameters in the local motion region are different from those in the still region. By making use of this fact, an approximate local motion region can be flagged. It is helpful to estimate accurately flicker parameter.2. Based on smoothness difference between two successive frames, a partial differential equation method for removing intensity flicker is developed. The method use the idea of minimization of total variation of an image. As a result, the method is noise-tolerable. At the same time, the order estimation of the flicker function can be avoided. Additonally, the smoothness and continuity of the flicker function can be satisfied as well.
Keywords/Search Tags:Digital Video Inpainting, Flicker Removal, Blotch Removal, Total Variation, Image Inpainting
PDF Full Text Request
Related items