摘要
基于ROI自适应的JPEG2000图像分区域压缩就是通过机器学习来自动确定ROI的生成,然后分别对感兴趣区域和背景区域进行不同压缩率的压缩,得到感兴趣区域清晰且数据量小的图片。分别介绍所用方法的原理和流程,并给出实现过程以及实验结果评估。实现图像分区域压缩这一功能,并给出SVM检测效果统计,JPEG2000分区域压缩和完全无损压缩的数据统计,证明这一程序的可行性和实用性。
ROI generated automatically based on JPEG2000 image compression in the different regions is to automatically determine the ROI generated by machine learning, then the region of interest and the background area is compressed at different compression rates, gets the picture with the clear ROI and a small amount of data. Introduces the principle and the method, and gives the process of implementation and evaluation of experimental results. Image compression and SVM performance statistics are achieved, JPEG2000 completely lossless compression and different regional statistics is showed, and it proves the feasibility and practicality of this program.
作者
孙佳全
SUN Jia-quan(College of Electronic Information Engineering,Tongji University,Shanghai 201804)
出处
《现代计算机》
2019年第6期72-78,共7页
Modern Computer
