Block-Based Reduced Resolution Inter Frame Coding with Template Matching Prediction

Abstract

A template matching prediction based on a group of reconstructed pixels surrounding a target block enables prediction of pixels in the target block without motion information. In this method, the predictor of a target block is produced by minimizing the matching error of the template. If the template contains features such as complex texture, an appropriate predictor can be produced. In flat regions, on the other hand, the residuals of a target block may become large due to the freedom possessed by the template. This paper proposes to predictively encode the decimated version of a target block in flat regions. Because of the high similarity between a target block and its decimated version in flat regions, coding bits for residuals of the target block can be reduced with little loss of quality. Simulation results show that the proposed scheme further improves coding efficiency of H.264 up to 6.5%.