Multimedia application has become extremely popular nowadays. Digital video products such as DVD player, DVD recorder, Digital TV system and Personal Video Recorder are entering everyone's home and bringing an exciting multimedia environment to the consumers. The improvement in the video quality is a key factor in the success of multimedia industry, which is facilitated by the advancement in multimedia, VLSI and networking technology. With the improvement in VLSI technology, more advanced and computationally complex video processing algorithms can be realized into consumer products and hence, producing high quality multimedia applications....[ Read more ]

Multimedia application has become extremely popular nowadays. Digital video products such as DVD player, DVD recorder, Digital TV system and Personal Video Recorder are entering everyone's home and bringing an exciting multimedia environment to the consumers. The improvement in the video quality is a key factor in the success of multimedia industry, which is facilitated by the advancement in multimedia, VLSI and networking technology. With the improvement in VLSI technology, more advanced and computationally complex video processing algorithms can be realized into consumer products and hence, producing high quality multimedia applications.

Postprocessing is the technique to enhance the video quality for display. In this thesis, two topics in postprocessing are investigated: error concealment and video format conversion.

Nowadays, many multimedia applications are integrated into some existing networks, for examples, video conferencing on the Internet, VOD with your 3G mobile phone, etc... However, those networks are unreliable and packets may be lost or corrupted during the transmission. The loss of information caused by the transmission errors can produce an adverse degradation in visual quality. Error concealment is a technique to recover the lost information and improve the quality of the corrupted video. In this thesis, we provide two algorithms to address the issue: "Temporal error concealment for video transmission with edge reconstruction" and "A Partitioned Linear Minimum Mean Square Estimator for Error Concealment". In the first algorithm, the edge information is considered in the estimation of the missing motion vector to improve the prediction accuracy. In the second algorithm, the multi-hypothesis motion compensated technique is employed to reduce the prediction error.

The second problem in multimedia application is the different display formats in use. For example, the TV transmission system (PAL, NTSC, SECAM) adopts interlaced video while the PC community and internet-broadcast adopts progressive video format. In order to provide interoperability between different systems, postprocessing is needed for conversion between different video formats. In particular, we address the problem of deinterlacing, which convert interlaced video to progressive video for display. In our spatial-temporal deinterlacing algorithm, it estimates the motion of each block and fills the missing field by motion compensation. If the estimated motion is not reliable, our spatial deinterlacing algorithm is invoked, which analyzes the content of a field and choose the most appropriate deinterlacing scheme. In addition to deinterlacing, our frame rate up-conversion algorithm can double the frame rate by interpolating the frame along the motion trajectory.