2026-11-092026-11-09https://scholars.csmu.edu.tw/handle/123456789/38756This research project concentrates on techniques including image representation and content-based image retrieval (CBIR) for liver tumor magnetic resonance (MR) images. In this project, the emphasis is on the development of an efficient and practical database system for recognizing and retrieving patterns in medical images that represent pathological processes. The goal is to present the physician with a selection of similar cases with known diagnoses. Content-Based Image Retrieval has been one of the most vivid research areas in the field of computer vision over the past 20 years. Recently content-based access to medical images for supporting clinical decision-making, telediagnosis, and education in medical schools has been proposed. We will develop a medical image representation structure, which can capture liver tumor shapes, sizes, and spatial relationships (including topological relationships and relative orientations of objects) between canonical organs and tumors. These features are important for disease diagnoses. Then we will establish a medical image database system with automatic indexing and content-based access capabilities for liver tumor MR images. We believe that this project will have significant contribution to the medical society in terms of medical image databases and disease diagnoses.本研究計劃就醫學影像之表示法與影像資料庫檢索自動化等技術，做深入的探 討，並提出有效的解決方案以建構成一個實際有用的整合系統，使能自動從醫學 影像資料庫中存取出相似病徵之圖片以輔助疾病診療。目的在於使醫師們能夠利 用這些已知的病例及有效的診療方法，來幫助有類似病徵的新病例診斷及醫療規 劃。以影像內容為搜尋基礎之影像資料庫擷取系統(Content-Based Image Retrieval, CBIR)，近二十年來更是熱門的研究主題，其是以影像視覺內容為基礎之影像搜 尋與擷取方法，這與以文字為搜尋對像的傳統式資料庫系統有明顯的不同。而影 像檢索自動化 (automatic image indexing)則是影像資料庫效能表現之關鍵技術。 最近十年來，該技術已陸續被應用到醫學影像，以搜尋相似病例輔助醫師看診、 遠距醫療(telediagnosis)、及作為醫學院相關教學之用。因此，本計畫主要目的為 以醫學影像中肝部腫瘤之形狀、大小及主要器官與腫瘤之空間關係做為搜尋相關 影像的特徵依據，而建構一有效率之醫學影像表示法，進而設計一個具有自動檢 索功能的肝部腫瘤影像資料庫系統，最後將作出實驗數據以評估其效果。相信本 研究計劃對醫學影像資料庫之研究與實際應用皆將會有重大之貢獻。醫學影像資料庫 影像表示法類似影像擷取自動檢索Content-basedMedical image databaseImage representationAutomatic indexingProject