頁籤選單縮合
題 名 | Color Image Segmentation for Identification of the Bladder Cancer |
---|---|
作 者 | Lee, Jiann-der; Hsiao, Yu-lin; Huang, Zhong-xian; | 書刊名 | 醫學工程 |
卷 期 | 11:3 民88.06 |
頁 次 | 頁149-157 |
分類號 | 410.1644 |
關鍵詞 | Image segmentation; Bladder cancer; |
語 文 | 英文(English) |
英文摘要 | Image segmentation is an essential step of many applications in the area of image processing, which entails the division or separation of the image into regions of similar attribute such as gray level, color, texture, etc. Due to the inherent noise and degradation, meaningful segmentation is not easy to obtain. In this paper, a novel algorithm is proposed to segment bladder cancer from the color image. This method first reduces a color image into a gray-level image and segments the image by thresholding. A morphology operation is then used to extract the initial border of cancer region. Next, an active contour model is employed to refine the initial boundary, by shrinking or expanding the boundary to approximate the actual border of the cancer region. In order to evaluate the performance of this algorithm, three evaluation functions are employed for comparison with other algorithms based on clustering techniques and thresholding techniques. These algorithms include of Spherical Coordinate Transform/Center 2-D Split and Principal Components Transform/median 3D split. The experimental results show that the proposed method is superior to the others both in processing time and accuracy. |
本系統中英文摘要資訊取自各篇刊載內容。