摘要
针对语义分割图像在存储和传输过程中压缩性能不足,占用资源多的问题,提出基于上下文与二进制算术编码的自适应编码方法。该方法提出了最佳上下文模型阶数的概念和计算方法,并根据最佳上下文模型与图像复杂度的相关性,设计了上下文模型自适应选择算法。通过支持向量机方法建立上下文模型分类器,根据语义分割图像复杂度特征,预测其最佳的上下文模型用于二进制算术编码。利用遥感影像语义分割数据集对算法进行测试,实验结果表明,与常用的二进制算术编码方法比,提出的算法能在压缩比不变的情况下缩短压缩时间38%左右,有效提高了压缩性能。
Aiming at the problems of insufficient compression performance and large resource consumption during the storage and transmission of semantic segmented images,an adaptive coding method based on context and binary arithmetic coding was proposed.This method put forward the concept and calculation method of the optimal context model order.Based on the correlation between optimal context model and image complexity,a context model adaptive selection algorithm was designed.A context model classifier is established by the support vector machine method,and image complexity feature was segmented according to semanticsto predict the best context model for binary arithmetic coding.The algorithm was tested by using the remote sensing image semantic segmentation dataset.Experimental results showed that the proposed algorithm could shorten the compression time by about 38%without changing the compression ratio in comparison with the commonly used binary arithmetic coding method,which effectively improved compression performance.
作者
陈鸿翔
梁晨光
李蒙
宫久路
CHEN Hongxiang;LIANG Chenguang;LI Meng;GONG Jiulu(School of Mechatronical Engineering, Beijing Institute of Technology, Beijing 100081, China;Beijing Institute of Aerospace Systems Engineering, Beijing 100076, China;Beijing Institute of Aerospace Control Devices, Beijing 100039, China)
出处
《探测与控制学报》
CSCD
北大核心
2020年第5期55-62,共8页
Journal of Detection & Control
关键词
语义分割
图像压缩
支持向量机
算术编码
semantic segmentation
image compression
support vector machine
arithmetic coding
作者简介
陈鸿翔(1995—),男,湖北襄阳人,硕士研究生,研究方向信号处理,图像编码,E-mail:chxhyfx@163.com。
