基于直线方向搜索的光线空间数据压缩方法

Ray-space data compression method based on rectilinear direction search

为了提高光线空间数据的压缩效率,提出了一种基于直线方向搜索的数据压缩方法.基于光线空间描述,讨论了其数据相关性和数据分布特性.在光线空间的数据压缩中,计算了数据的相关系数,采用二维离散余弦变换(DCT)进行了变换域分析.在片间预测中,通过对光线空间片的纹理统计,确立初始的搜索方向,实现分层的直线方向搜索方法.以复杂度失真最优化作为层间判别准则,通过引入自适应阈值实现块匹配.实验结果表明,与全搜索方法相比,该方法计算复杂度明显降低,同时保持了近似的解码图像质量和编码码率,进一步提高光线空间数据的编码效率.

To improve the efficiency of ray-space data compression, a data compression method based on rectilinear direction search was proposed. After the information ray-space was described, the data correlations and distribution characteristics were discussed. Data correlation coefficients were calculated and switch compre regions were analyzed by using two dimension discrete cosine transform （DCT） in ray-space data ssion. By adopting texture statistics for ray-space slice, initial search direction was established, and then multi-layer rectilinear direction search method was realized in inter-slice prediction. Optimization of complexity and distortion was taken as inter-layer discriminant criteria, and block matching was achieved by introducing self-adapting threshold. Experimental results show that the proposed method can reduce the computational complexity significantly compared with the full search method, and maintain similar decode visual quality and coding bit rate, then improve the ray-space data coding efficiency.