Non-subsampled Contourlet Based Deformable Model For Medical Image Segmentation

Artifacts involved in medical images like noise and intensity inhomogeneity makes an automated approach of segmentation failure. This paper proposed deformable model based automatic image segmentation model with above said constraints. The deformable models are sensitive to the noise and image intensities. Therefore use of non-subsampled contourlet transform is proposed to tackle these artifacts and to enhance images. These enhanced images are utilized along with region based deformable model for successful segmentation. The illustrated results on various types of medical images show the advantages of proposed model.