Automatic detection and restoration of frame pixel-shift in videos

作者：Yuan, Haidong; Ma, Huadong^*; Huang, Xiaodong

来源：Multimedia Tools and Applications, 2010, 47(2): 307-323.

DOI：10.1007/s11042-009-0324-6

• Pixel-shift
• Defect detection and restoration
• Content management

Defect detection and restoration of degraded videos is an important topic in media content management systems. Frame pixel-shift is a common form of severe defect in videos caused by loss of consecutive pixels by the video transmission system. Pixel-shift refers to the large amount of pixel shifts one by one due to a small quantity of image data loss. The damaged region in the affected frame is usually quite large, causing serious degradation of visual quality. This paper addresses the issue of how to automatically detect and restore frame pixel-shift in videos. Pixel-shift frame detection relies on spatio-temporal information and motion estimation. Accurate measurement of pixels shift is achieved based on the analysis of temporal frequency information and restoration is accomplished by reversing the pixels shift and spatio-temporal interpolation. Performance evaluation using real video sequences demonstrate the good performance of our algorithm.

• 出版日期2010-4
• 单位北京邮电大学; 中国人民解放军信息工程大学