Wireless channels are typically noisy and suffer from a number of channel degradations such as bit errors and burst errors due to fading and multipath reflections. Video compression techniques typically apply predictive coding, that is, code the information in the current frame of the video signal predictively as a difference signal with respect to the previous frame. They also employ variable-length coding schemes such as Huffman codes to achieve a further degree of compaction. The compressed video signal is extremely vulnerable to transmission errors. In INTER mode, the loss of information in one frame has considerable impact on the quality of the following frames. As a result temporal error propagation continues for a longer period of time. The error may also propagate spatially from its original occurrence due to motion-compensated prediction. For this reason, various concealment techniques must be applied by encoder, decoder and both encoder and decoder to reduce the severity of errors.
The proposed RPSS (Reference Picture Segment Selection) mode is the RPS mode in picture segment of the ISD mode. The RPS mode has the weakness that it has to change the whole reference picture even if transmission error occurs only in the small part of picture. If the RPS mode is applied only to an error-occurred part and a conventional coding scheme is applied to an error-free part, the higher coding efficiency can be obtained. Simulation results show that there is the optimal number of segment which provides the best performance with given environment such as the BER (Bit Error Rate) of wireless channels, the characteristics of input sequence, what kind of RPSS mode is used, etc. To apply the channel-adaptive RPSS mode later, the scheme which can change the number of segment without INTRA refresh and spacial error propagation is proposed in this thesis.
