A perception-motivated image interpolation algorithm

Image interpolation, or to obtain a high-resolution image from a corresponding low-resolution image, is still a hard question. In order to better solve this question, we demonstrate a partitioned image interpolation model and propose a perception-motivated image interpolation algorithm according to human eye visual mechanism (PMIA). The PMIA main implementation includes two processes. Firstly the original image is divided into the attention region, the transition region and the general region and then different interpolation mode can be used based on different regions. Conducted experiments have shown that our algorithm can spend less time to produce the most satisfactory image. From the perspective of engineering application, our method can reduce the communication bandwidth and time consuming.