Geometric Active Contour Models Applications And Research In The Image Of The Target Contour Segmentation

According to different image analysis task,image segmentation aims at partition the original image into several meaningful regions or isolating the region of interesting(ROI).At present, the main image segmentation algorithms are splited into based on edges and based on regions. The paper's main task is concentrated on the former.At present,active contour model has become a important tool in image analysis field.Nowadays,active contour model has been divided into two classes,parametric and geometric active contours. This paper analyzes active contour model' principles methods and it's applications,and focus on geometric active contours in ultrasonic medicine and weak-edge regions and color image segementation' application. The paper's specific content include following three parts.Firstly,advanced a improved fast method in image segementation according to ultrasonic image's characteristic. The method have following advantages :firstly,it can weighing the relations between denoising and edge-preserving,duo to total variational method that have the characters of denosing and preserving image's edge;secondly,this method has fast image segmentation speed,according to the compute complexity of level set,the paper advanced a narrow band algorithm of level set which can greatly decease compute complexity and largly raising curve's evolve speed;thirdly,it can combine with level set to segmentation muti-objects simultaneous.Secondly, the paper advanced a image segmentation method aimed at weak-edge regions based on Mean Shift cluster segementation. This method have following advantages:firstly,it can resistant to noise effectively; secondly,this method can segment weak-edge regions accurately;thirdly,it can combine with level set to handle curves' topology freely that can segmentation muti-objects simultaneously.Thirdly,the paper advanced a colour image segment model based on color gradient model aimed at segment colour objects. the model have following advantages:firstly,it has robust performance that duo to color gradient most reflect the object's material chang in color image segementation;secondly,it have the advantages of utility and larger application range that the model robust to color space in image processing field;Finally,this paper reviewed whole outline and content,pointed out the insufficiencies and make a future outlook in my paper.