A New Hybrid Classification-Based Adaptive Error Concealment Technique for Image Transmission through Wireless Communication Channels

摘要

In image communications, error concealment is an effective technique for reconstruction of some of the lost contents in image or video during transmission in practical wireless communication channels. There are many known techniques for error concealment; each one gives a different degree of image quality. Thus, there is no preferred technique all the time because some of these techniques give a good objective image quality while others give a good subjective image quality. Multiple techniques can be used by making an image adaptive classification in transmitter, or receiver, to select the suitable technique for each block in the image. In this way, one uses the advantages of each technique to obtain better overall objective and subjective image quality. In this paper, two algorithms for error concealment are considered, namely, the Orientation-Adaptive-Sequential-Interpolation (OASI) technique and the Long-Range-Correlation (LRC) technique. The approach is to adaptively combine the two mentioned algorithms to get better overall performance. Numerical examples show that the suggested hybrid technique results in improved performance over different examples of communication channel impairments including fading and additive noise. Combination of the suggested error concealment technique with channel coding including convolution coding and block interleaving is also considered.