Motion JPEG2000 Coding Scheme Based on Human Visual System for Digital Cinema

摘要

In this paper, the properties of human visual system such as insensitivity to high temporal frequency and contrast masking are exploited to improve the subjective coding efficiency of the Motion JPEG2000. Specifically, we detect motion information using DWT (Discrete Wavelet Transform) coefficients of the JPEG2000 and then determine the importance of the corresponding high frequency components based on contrast masking effect. If the motion turns out to be imperceptible by the human vision system, then we discard the corresponding subband data. Proposed scheme yields higher subjective quality at lower bit-rate than the conventional Motion JPEG2000 coding scheme. Also, the proposed algorithm needs lower memory access and complexity than the previous enhanced coding scheme using motion estimation.