透過您的圖書館登入
IP:18.117.81.240
  • 期刊

加強式中心映射法於相似醫學影像集壓縮之研究

Enhanced Centroid Method for Compressing Similar Medical Images

摘要


醫學影像所擁有的資訊,是醫生作為病情判斷一個重要的依據,其多為一組資料量龐大的相似影像群集合。所謂相似影像群是指具有相似的特徵、像素分佈以及邊緣分佈之影像集合。換句話說,集合中影像與影像之間具有某種程度的相似性(Similarities)。因此相較於單張影像,相似影像群具有「集合重複性」的特性。本論文採用加強式壓縮模型作為壓縮相似影像集合的基本架構,這個模型乃增加了減少集合重複性的步驟,稱之為集合重複性壓縮法。集合重複性壓縮法分為極小極大差異法與極小極大預測法,以及中心法。本論文使用中心法產生N張灰階相似影像集合的平均影像來作為相似影像集合的共通圖像,並對個別影像擷取出差異值做小波壓縮,並與一般單一影像壓縮做壓縮率比較。實驗結果顯示,應用小波轉換於中心映射法對於連續相似影像集合在壓縮率上比僅使用熵編碼(Entropy Coding)提高約56%、比使用小波壓縮編碼(Wavelet Compression)提高約26%,證明減少集合重複性的確能改善連續相似影像集合壓縮之效能。

並列摘要


Similar images are images with common features, similar pixel distributions, and similar edge distributions. Fields such as medical imaging often need to store large collections of similar images. In a set of similar images the images similarities represent patterns that consistently appear across all images; this results in ”set redundancy”. Recently, the research of wavelet transformation is developed quickly and compression using wavelet is a good choice instead of DCT transform. In this paper, we present the Centroid method, which extracts the similarity coefficients obtained from wavelet transform, to reduce set redundancy and achieve higher compression ratio for sets of similar images. Experimental results with a set of CT images demonstrate that the Centroid method using wavelet can deliver significantly improved image compression. Compared with entropy coding and wavelet compression respectively, the proposed method is improved about 56% and 26%.

延伸閱讀