Lossless Video Coding Based on Multi-Frame Example Search and Probability Model Optimization
Lossless Video Coding Based on Multi-Frame Example Search and Probability Model Optimization
复制标题
基于多帧实例搜索和概率模型优化的无损视频编码
DOI：
10.1117/12.2521541
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
Susumu Itoh
中科院分区：
文献类型：
--
作者：
Koji Nemoto;Yusuke Kameda;Ichiro Matsuda;Susumu Itoh
本文描述了一种新的无损视频编码方法，该方法直接估计逐个像素的图像值的概率分布。在估计过程中，从位于当前帧以及先前帧的已编码区域上的搜索窗口收集几个示例，即其邻域类似于要编码的目标像素的局部纹理的像素集合。然后将概率分布建模为中心位置由个别实例给出的高斯函数的加权和。此外，对控制高斯函数形状的模型参数进行了数值优化，使得所得到的编码率可以是最小的。仿真结果表明，增加参考帧数目可以提高编码性能。
This paper describes a novel lossless video coding method that directly estimates a probability distribution of image values pel-by-pel. In the estimation process, several examples, i.e. a set of pels whose neighborhoods are similar to a local texture of the target pel to be encoded, are gathered from search windows located on an already encoded area of the current frame as well as those of the previous frames. Then the probability distribution is modeled as a weighted sum of the Gaussian functions whose center positions are given by the individual examples. Furthermore, model parameters that control shapes of the Gaussian functions are numerically optimized so that the resulting coding rate can be a minimum. Simulation results indicate that the coding performance can be improved by increasing the number of reference frames.