講演抄録/キーワード |
講演名 |
2008-01-25 11:30
A Rule Based Technique for Liver Segmentation in CT Data Amir H. Foruzan・○Reza A. Zoroofi(Univ. of Tehran)・Yoshinobu Sato・Masatoshi Hori(Osaka Univ.) MI2007-71 |
抄録 |
(和) |
Segmentation is a major step in liver image analysis and may help clinician to quantify the progression of liver diseases. A physician performs liver manual segmentation based on anatomical facts. We follow a similar approach in the proposed techniques. We first find the image slice corresponding to the largest cross-section of the liver. We regard it as the starting slice in segmentation. Using this slice, we estimate the intensity range of liver and apply the range used throughout the algorithm. We next extract bones and connect them to form a mask for liver region. By this mask, we remove some non-liver parts within the intensity range of liver. We then utilize several anatomical facts to convert gray-scale liver images to colored ones. Results reveal that the proposed technique outperforms other existing techniques in terms of speed, accuracy, sensitivity, and specificity. |
(英) |
Segmentation is a major step in liver image analysis and may help clinician to quantify the progression of liver diseases. A physician performs liver manual segmentation based on anatomical facts. We follow a similar approach in the proposed techniques. We first find the image slice corresponding to the largest cross-section of the liver. We regard it as the starting slice in segmentation. Using this slice, we estimate the intensity range of liver and apply the range used throughout the algorithm. We next extract bones and connect them to form a mask for liver region. By this mask, we remove some non-liver parts within the intensity range of liver. We then utilize several anatomical facts to convert gray-scale liver images to colored ones. Results reveal that the proposed technique outperforms other existing techniques in terms of speed, accuracy, sensitivity, and specificity. |
キーワード |
(和) |
Liver boundary detection / Knowledge based segmentation / Liver segmentation / Adaptive Thresholding / Color map generation / / / |
(英) |
Liver boundary detection / Knowledge based segmentation / Liver segmentation / Adaptive Thresholding / Color map generation / / / |
文献情報 |
信学技報, vol. 107, no. 461, MI2007-71, pp. 47-54, 2008年1月. |
資料番号 |
MI2007-71 |
発行日 |
2008-01-18 (MI) |
ISSN |
Print edition: ISSN 0913-5685 Online edition: ISSN 2432-6380 |
著作権に ついて |
技術研究報告に掲載された論文の著作権は電子情報通信学会に帰属します.(許諾番号:10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
PDFダウンロード |
MI2007-71 |