朴 賢敬, 西村 敏博

Hyunkyung PARK, Toshihiro NISHIMURA

早稲田大学大学院情報生産システム研究科

Graduate School of Information, Production and Systems，Waseda University

キーワード :

1. Introduction
Recently, medical image equipment and processing technology are development occur to medical diagnosis field. Most of the analysis, such as surgical planning, simulation, diagnosis and therapy evaluation depends on the results of the segmentation procedure. However, the medical and anatomical specialist is a possible variously discrimination about the same medical images which is differ to the result sometimes. That reason, it is necessary to use the computer image processing or automation treatment. Then we proposed the based on threshold level set method by ultrasound images segmentation.

2. The threshold Level Set method
The Level set formulation was first proposed as a numerical technique that tracks and evolving contour. It originates from the theory of propagating solid/liquid interface with curvature dependent speed. The main idea is to represent a curve as the zero level set of a higher dimensional function with the motion of the curve embedded in the motion of the higher dimensional surface.
We proposed method is consists of decision of the weighted length through histogram analysis supposed to set threshold value. The coefficient of weighted area term is decided for threshold value. If the initial contour is outside the object, the coefficient of weighted area term chooses a positive. All of the pixel intensities are calculated mean value. That will determine the threshold value and parameter of the weighted area term. We can be determining to three thresholds

3. Results and Conclusion
We have extract to our method better than the original level set method real medical ultrasound images contour decision. With adjust the threshold by user, we will be get desired results by weighted area term parameter.