基于深度学习CNN的图像压缩编码技术研究  

Research on Image Compression Coding Technology Based on Deep Learning CNN

作　　者：吴莹莹 彭亚雄 陆安江 WU Yingying;PENG Yaxiong;LU Anjiang(Guizhou University,School of Big Data and Information Engineering,Guiyang Guizhou 550025)

机构地区：[1]贵州大学大数据与信息工程学院,贵州贵阳550025

出　　处：《软件》2020年第12期18-23,共6页Software

基　　金：贵州省科技重大专项:草海综合整治工程大数据系统集成与示范(黔科合重大专项字[2016]3022号);贵州省科技成果转化项目:可溯源的天然饮用水电子商务系统平台研发(黔科合成果[2017]4856)。

摘　　要：随着信息通信技术的发展以及智能数码设备的普及,为了可以更便捷获取高清海量的图片信息,在存储空间有限的情况下,信号和图像需要进行压缩后传输以减少所占用的存储空间。而深度学习在图像处理领域方面是现今研究的热点,本文旨在以卷积神经网络在图像压缩编码应用中减小传输数据的规模,实现图像信息的高效传输,对图像特征提取,量化,熵编码以及图像质量评价标准等方面进行现状分析与研究。With the development of information and communication technology and the popularization of smart digital devices,in order to more conveniently obtain high-definition mass picture information,in the case of limited storage space,signals and images need to be compressed and transmitted to reduce the occupied storage space.Deep learning is a hot research topic in the field of image processing.This article aims to reduce the scale of transmission data in the application of image compression and coding with convolutional neural network,realize the efficient transmission of image information,extract image features,quantify,and entropy Coding and image quality evaluation standards and other aspects of the current situation analysis and research.

关 键 词：图像压缩 卷积神经网络 量化 熵编码 

分 类 号：TN911.73[电子电信—通信与信息系统]