PROJECT TITLE :

MLP neural network classifier for medical Image Segmentation - 2016

ABSTRACT:

The selection of a segmentation methodology depends on several concerns, particularly the character of the image, the primitives to extract and therefore the segmentation ways. We have a tendency to propose an MLP-basis neuronal approach for the selection of the segmentation methodology taking into consideration the character of the input image. First, an analysis of the standard of segmentation by completely different ways and using numerous criteria of evaluation was distributed. Then, a characterization of pictures, based on some objective parameters, was performed. The ensuing descriptors will be used as input to the neuronal approach to associate each type of image with the adequate segmentation methodology when learning. We report the results of the intelligent segmentation methodology choice obtained on totally different databases of medical images. The discussion of those encouraging results allowed us to boost our success rate and cowl all styles of images.