On edge detection in MRI using the wavelet transform and unsupervised neural networks

This paper investigates a novel feature extraction approach to MRI edge detection based on identifying the critical image edges by formulating the problem as a two-stage unsupervised classification task. The main goal of such a research effort is to better identify abrupt image changes without increasing the presence of noise in the resulting image. The suggested methodology is based on novel wavelet descriptors involving the discrete k-level 2-D wavelet transform applied to sliding windows raster scanning the original image as well as on vector quantizing self-organizing feature maps (SOFM) and SVD analysis. This edge detection process is considered as a two-stage clustering procedure employing SOFM trained with the Kohonen algorithm. The feasibility of this novel proposed approach is studied by applying it to the edge detection structure segmentation problem of brain slice MRI images.