請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/74734
標題: | 應用於心音訊號壓縮的卷積自動編碼器 Heart sound Compression based on Convolutional Autoencoder |
作者: | Kai-Chieh Hsu 許凱傑 |
指導教授: | 曹恆偉(Hen-Wai Tsao),錢膺仁(Ying-Ren Chien) |
關鍵字: | 訊號壓縮,心音圖,深度學習,自動編碼器,遠端醫療, Signal compression,PCG,Deep Learning,Autoencoder,Telemedicine, |
出版年 : | 2019 |
學位: | 碩士 |
摘要: | 近年來智慧醫療領域越來越受重視,許多通/資訊科技包括雲端、物聯網、遠距、大數據分析及人工智慧等,已大量應用於醫療領域中。隨著社會人口年齡層的老化,慢性病的患者逐漸增加,應用於長期病患監測服務、行動照護的需求漸增。
心音圖提供一種非侵入式檢測方法,用來偵測心臟瓣膜異常及輔助心臟病的病因判斷,在長期心音圖檢測而產生龐大資料量情況下,一個有效率的訊號壓縮系統是必要的。目前的醫學檢測大都是在醫療院所完成,不少病患為了檢查及回診,需來回奔波,浪費病患等候看診的時間,也浪費不少社會資源。對此,發展遠距醫療(telemedicine)是台灣現今一個很重要的課題,遠距醫療是利用遠距通訊傳遞醫學資訊的一種新科技,更重要的是,開創了一種新的醫學溝通方式,使醫師與病人間可進行同步與非同步的互動,克服空間與時間的障礙,改善醫療品質、降低社會成本及增加便利性。本論文提出一種利用深度卷積自動編碼器的心音圖壓縮方法,對遠端通訊網路中干擾所造成的錯誤,有一定的容忍度,可應用於遠端醫療上,能改善發展遠距醫療會受到的限制。 本論文首章節為論文簡介,第二章及第三章為背景知識介紹,從第四章開始為本論文主要貢獻,也就是系統架構設計,包含心音訊號切割、心音特徵定義、自動編碼器架構…等,並在第五章探討模擬測試結果,第六章為結語與未來展望,最末章為附錄。 In the past few years, the field of smart health has been more and more important. Lots of technology such as cloud computing、internet of things、remote control、big data analysis and artificial intelligence has already applied to medical field. As the population ages, the population of chronic diseases gradually increases. The demand for long-term patient monitoring services and action care is increasing. The Phonocardiogram(PCG) provides a non-invasive method for detecting heart valve abnormalities and assisting in the diagnosis of heart disease. Due to huge amounts of data generated by long-term PCG monitoring, an efficient signal compression method is necessary. Most of the current medical tests are done in medical institutions. Many patients need to go back and forth in order to check and return to the hospital. This wastes the patient’s time for medical consultation, and also wastes a lot of social resources. Therefore, the development of telemedicine is a very important issue today. Telemedicine is a new technology that uses telematics to transmit medical information. More importantly, it has created a new way of medical communication, enabling synchronized and asynchronous interaction between physicians and patients, overcoming space and time barriers, improving medical quality, reducing social costs and increasing convenience. This paper proposes a PCG compression method using deep convolutional auto-encoder. This method has a certain allowable error rate for interference errors in remote communication, and can be applied to telemedicine, which will slove the problems of developing telemedicine. |
URI: | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/74734 |
DOI: | 10.6342/NTU201904432 |
全文授權: | 有償授權 |
顯示於系所單位: | 電信工程學研究所 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-108-1.pdf 目前未授權公開取用 | 3.98 MB | Adobe PDF |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。