Detection and Classification of MRI Brain Images For Head/Brain Injury Using Soft Computing Techniques
Burje Shrikant1,*, Prof. Dr. Rungta Sourabh1, Prof. Dr. Shukla Anupam2
1Rungta College of Engineering and Technology, Bhilai, India
2IIITM, Gwalior, M.P., India
*Corresponding Author E-mail: firstname.lastname@example.org
Online published on 29 April, 2017.
It is essential to have a rigorous computerized system for Magnetic Resonance Images (MRI) of the brain for medical perception and clinical analysis. This article focuses on our proposed method of hybrid approach for classification of normal and abnormalities in magnetic resonance brain images. Wavelet and PCA were functioning feature extraction and reduction from image respectively. The featured images finally were linked to Neuro-Fuzzy Classifier (NFC) for classification. The proposed methodology, including three basic steps, namely preprocessing, training and classified output. It extracts and reduced the dimension of features from the set of scan brain MR images of patients. Once preprocessing done, the featured image trained by soft computing based fuzzy neural tool and finally fed to the Neuro-Fuzzy Classifier (NFC) for detection of abnormalities in new MR images. The Hybrid NFC is combined with K-fold fuzzy C-means Neuro-Fuzzy Classifier which is used to enhance Abstraction of NFC. We focus on common brain diseases such as meningioma, Alzheimer's and visual agnosia as an abnormal brain. K-Fold Neuro Fuzzy Classifier provides the accuracy around 98% with minimum computational time.
MRI, PCA, NFC, DWT, PSNR.