This paper proposes a new approach for automated detection of multiple sclerosis lesions which separate the lesion region from healthy tissues in magnetic resonance images. The magnetic resonance feature images used for the detection consists of proton density (PD-) and T2-weighted brain MR images for each axial slice through the head.This method is used to filter false detections of small multiple sclerosis lesions in magnetic resonance images based on the analysis of regional features. In this method fuzzy rules are used to detect the image pixels showing hyperintensity. The sclerosis lesions are visualized using Magnetic resonance imaging (MRI). The analysis of the results obtained allows to extract some features with minor differences between small multiple sclerosis lesions and false detections. These features are used as restrictions for obtaining a new and improved fuzzy membership function associated with the presence of hyperintensity. Results show a reduction in number of false detections and preserves the small multiple sclerosis lesions detected.The application of the proposed method for early detection of multiple sclerosis is demonstrated to improve efficiency and accuracy of clinical practice.
Volume 11 | 01-Special Issue
Pages: 862-869