A Survey Report on Brain Image Segmentation

This paper describes various techniques for classification of the magnetic resonance human brain image. It includes hybrid techniques such as Automatic, Semi-automatic for feature extraction, dimensionality reduction and classification. Feature extraction and reduction using various mathematical techniques such as Discrete Wavelet Transform, Principal Component Analysis.Various classifiers are used based on Back propagation Network FP-ANN and K-nearest neighbor. Numeric data or multimodal data can be classified using Optimal Semi-supervised Fuzzy C-means (FCM) and Probability Density Function (PDF) estimation. Different papers present method for semi-supervised Maximum a Posterior segmentation of brain tissues. There are two types of segmentation-Supervised and Unsupervised. Segmentation of Magnetic Resonance (MR) images is a process of delineation of regions representing different types of tissues. Finally all the techniques are compared on the basis of feature extraction techniques used and reduction of dimensionality and classification methods. Thus this paper provides a complete literature survey of medical imaging analysis based on segmentation.