| 講演抄録／キーワード |
| 講演名 |
2007-01-26 16:35
A Hybrid Technique for Automatic Quantification of Hip Joint Cartilages in MR Images
M. Khanmohammadi・○R. A. Zoroofi（Univ. of Tehran）・Y. Sato・T. Nishii・K. Nakanishi・H. Tanaka・N. Sugano・H. Yoshikawa・H. Nakamura・S. Tamura（Osaka Univ.） |
| 抄録 |
（和） |
The estimation of hip joint cartilage’s thickness is clinically important. Automatic segmentation of the hip cartilage is the first step in volume rendering of the cartilage as they can not visualize directly from original images. In this research, we propose an automatic multi step technique for segmentation the femoral head and acetabulum using clinically obtained multi-slice T1-weighted MR data. First we resample data sets with a modified Sinc interpolation technique to perform an isotropic dataset. Using anatomical knowledge of the femur, we estimate the center of the femoral head by a Hough transform. We develop a multistage approach for bone segmentation that is based on vector quantization technique and a region-growing strategy. We localize the hip joint space edges by customizing the anatomical constraint. Then, from our recent research we use localized articular space central line to segment the acetabular and femoral head cartilages from each other. After segmentation, we compare two methods for thickness measurement that are based on extremum points found from images enhanced by directional derivative filters, and the distance between the edge points along iso normal of each point. Finally we visualize the cartilages applying the maximum intensity projection as one of the volume ray casting approaches. The techniques are implemented in C++ and Matlab programming languages. The feasibility of the proposed techniques is successfully evaluated in the presence of 40 hips including 1200 MR images. |
| （英） |
The estimation of hip joint cartilage’s thickness is clinically important. Automatic segmentation of the hip cartilage is the first step in volume rendering of the cartilage as they can not visualize directly from original images. In this research, we propose an automatic multi step technique for segmentation the femoral head and acetabulum using clinically obtained multi-slice T1-weighted MR data. First we resample data sets with a modified Sinc interpolation technique to perform an isotropic dataset. Using anatomical knowledge of the femur, we estimate the center of the femoral head by a Hough transform. We develop a multistage approach for bone segmentation that is based on vector quantization technique and a region-growing strategy. We localize the hip joint space edges by customizing the anatomical constraint. Then, from our recent research we use localized articular space central line to segment the acetabular and femoral head cartilages from each other. After segmentation, we compare two methods for thickness measurement that are based on extremum points found from images enhanced by directional derivative filters, and the distance between the edge points along iso normal of each point. Finally we visualize the cartilages applying the maximum intensity projection as one of the volume ray casting approaches. The techniques are implemented in C++ and Matlab programming languages. The feasibility of the proposed techniques is successfully evaluated in the presence of 40 hips including 1200 MR images. |
| キーワード |
（和） |
Cartilage Quantification / Directional Filter / Vector Quantization / Bone Segmentation / / / / |
| （英） |
Cartilage Quantification / Directional Filter / Vector Quantization / Bone Segmentation / / / / |
| 文献情報 |
信学技報, vol. 106, no. 509, MI2006-126, pp. 161-164, 2007年1月. |
| 資料番号 |
MI2006-126 |
| 発行日 |
2007-01-19 (MI) |
| ISSN |
Print edition: ISSN 0913-5685 |
| PDFダウンロード |
|