面向医疗文本数据压缩的主流算法及发展趋势  

Summary of mainstream algorithms and development tendency for medical text data compression

作　　者：高克承 徐桓 刘岩 刘洋 张曦 GAO Ke-cheng;XU Huan;LIU Yan(School of Biomedical Engineering,Air Force Medical University,Xi’an 710032,China;不详)

机构地区：[1]空军军医大学军事生物医学工程学系,陕西西安710032 [2]联勤保障部队药品仪器监督检验总站,北京100071

出　　处：《中国医学装备》2020年第11期195-199,共5页China Medical Equipment

基　　金：国家自然科学基金面上项目(81871424)“基于多模态MR影像的胶质母细胞瘤高危区域定位及预后预测研究”;国家自然科学基金青年科学基金(81701658)“基于多模态MRI与卷积神经网络的较低级别胶质瘤IDH突变预测研究”。

摘　　要：随着远程医疗和网络诊断的出现,医疗卫生数据及患者个人信息量呈海量化,为了解决医疗卫生数据存储与传输困难的问题,针对以往经典压缩算法及目前的压缩算法新进展,介绍适用于医疗卫生的文本数据压缩算法。以往经典的压缩算法存在压缩率不足及通常适用于小文件的缺点,对此相关研究提出了适用于海量文本并具有高压缩率的压缩感知算法。通过对机器学习和深度学习用于文本压缩进行相关讨论,为日后文本压缩发展研究提供参考。In recent years,with the emergence of telemedicine and network diagnosis,the medical health data and patient’s personal information grew larger.In order to solve the current problems of difficult storage and transmission of medical health data,text data compression algorithms that might be applied to health and medical care were introduced and elaborated as the past classic compression algorithms and new developments in current compression algorithms.Due to the shortcomings of the past classical compression algorithms with insufficient compression rate and that was usually applied to small files,the correlational research literature proposed compression-sensing algorithms that were suitable for mass text and had high compression rates.At the same time,the paper provided reference for the research of the development of text compression in the future through the relevant discussion of machine learning and in-depth learning that were applicable to text compression.

关 键 词：医疗卫生数据压缩 医疗文本 HUFFMAN编码 LZW编码 压缩感知 

分 类 号：R-058[医药卫生]