Designing and implementing mutual scan-line fractal coding for field-sequential stereo video sequences

Fractal coding is a recent and effective compression technique based on self-similarity of images. However, its encoding time is very high. In addition, when 2-D fractal coding is applied directly to field-sequential stereo video sequences, the high vertical-frequencies of stereo image pairs are suppressed. This makes the depth perception defective. To overcome these problems, we propose a mutual scan-line fractal coding (MSFC). In this method, there are two significant issues that differ from traditional fractal coding. The first is image partition; the scan-line approach is used to divide the range blocks of images for preserving the high vertical-frequencies. The second is domain pool design; the mutual domain pools based on equivalent property of two complete metric spaces are designed for speeding up the encoding and decoding times. The experimental results illustrate that the proposed approach can improve image quality in terms of preserving high vertical-frequencies, reduce both encoding and decoding times, and maintain coding efficiency in bit rate at range of 1.20-2.00 bpp without any entropy encoding at the end of the scheme.