Multiresolution wavelet-fractal analysis of tumor detection |
|
|
The goal of this project is to develop an innovative approach which exploits the excellent feature localization property of multiresolution wavelet and fractal nature of abnormal lesions in fractional Brownian motion (fBm) framework that will provide a promising direction for tumor and hard-to-detect abnormalities identification in brain MR images. We investigate the multiresolution wavelet filter to obtain localized MR image features and use these features to design fractal filters in fractional-Brownian motion framework. The resulting filtered multiresolution wavelet-fractal features indicate the object of interest in different MR image sequences. We then fuse these multiresolution-fractal features to classify different types lesions in brain MR images. We purpose to develop a powerful Feature Extraction and Abnormality Identification (FEAI) engine to include multiresolution filter design, fractal filter design, multiresolution-fractal feature identification and feature fusion in the brain MR image sequences. The proposed fBm -based multiresolution approach will provide a radiologist, student or researcher with a feature-based reliable MR image recognition engine for aid to tumor and hard-to-detect abnormalities identification, learning and research respectively.
|
Publications
Contact Us
Website
|
