徐麗蘋Li-Pin Hsu葉宥徵Yu-Cheng Yeh醫學資訊學系碩士班2026-10-139999-12-312014https://handle.ncl.edu.tw/11296/khzax9https://scholars.csmu.edu.tw/handle/123456789/21175本論文的主要目的為提出一新的影像表示法，能將乳房鈣化點診斷的所有依據因素有效且整合性地表示出來，並且依此表示法提出能輔助醫師診斷的重要功能。本論文亦基於此影像表示法提出能自動化判斷鈣化點分布類型之演算法，有效地輔助醫師對於乳房鈣化點為良惡性之診斷的效率，並提昇診斷與治療規劃的精確度。另外，此影像表示法具有可視覺化的功能，使醫生不需從資料庫調出原圖便能快速瀏覽圖片的主要內容。本研究之實驗為採用實際的乳房攝影影像，實驗結果證明本論文所提的方法是有效且精確的。In this paper, a new image representation method for assisting the doctors in diagnosing breast calcifications as benign or malignant is proposed. This image representation method can represent all diagnostic factors effectively, and some important functions of assistance are also provided based on this image representation method. Based on this image representation, an algorithm to automatically identify the distribution type of calcifications is also provided, which can assist physicians to analysis breast calcifications efficiently as well as to make the correct diagnosis and therapy planning. Furthermore, visualization function based on this image representation can be easily built, which can help physicians with browsing mammographic images quickly. In this study, the actual clinical calcification imaging experiments have been performed. Experimental results show that the method proposed in this paper is effective and accurate.thesis