A New Image Segmentation Approach using Region-based Active Contour Model

YIN limin
2011 International Journal of Advancements in Computing Technology  
Over the Past decade Medical Image segmentation is one of the most challenging and focused topic for intensive research in interdisciplinary areas of Image processing and computer vision. Segmentation is the process of automatic or semi-automatic detection of boundaries [5] . In this paper, we implement a novel unsupervised method for segmenting MRI brain Images based on multiresolution transforms and region based active contour. Application of multiscale, multiresolution methods with active
more » ... hods with active contour is most interesting research topic in image segmentation [6] .This new application makes segmentation algorithms more economical for computation. Keywords: Multiscale and Multiresolution Transform, Chanvese active contour, Curvelet transform 2) The curve evolution speed is very slow, as a result convergence is also slow and the level set formulation requires reinitialization at every step during evolution. 3) AC method has a great sensitivity to noise which may yield false segmentation results. Meanwhile, for noise images, the gradient descent flow requires much more expensive computation and iterations to force the active curve(s) to converge.
doi:10.4156/ijact.vol3.issue10.43 fatcat:oiwohdtm2beezmohtnrx3fagca