Ms. R. J. Deshmukh et al Int. Journal of Engineering Research and Applications www.ijera.com ISSN : 2248-9622, Vol. 4, Issue 5( Version 4), May 2014, pp.78-80 www.ijera.com 78 | Page A Survey on Automated Techniques for Brain Tumor Detection Ms. R. J. Deshmukh*, Prof. R. S. Khule** *(Department of Electronics&Telecommuincation Engineering, Pune University, Nasik.) ** (Department of Electronics&Telecommuincation Engineering, Pune University, Nasik) ABSTRACT Brain tumor detection is difficult and complicated job for radiologist. Manual segmentation of brain tumor is tedious job and may provide inaccurate results. So there is a need of automated technique for accurate brain tumor detection. Different automatic methods have been developed till date to increase the accuracy for tumor diagnosis. This paper reviews research work on computer aided diagnosis (CAD) done by researchers. In review paper different methods of brain tumor detection used uptil now is summarized with merits and demerits of earlier proposed work. Keywords – MRI, Segmentation, MLP, Clustering. I. INTRODUCTION Brain cancer is most deadly diseases. Brain tumors are of two main types: (i) Benign tumors (ii) Malignant tumors. Benign tumors are noncancerous tumors and do not spread. Malignant tumors are typically called brain cancer contain cancerous cells and grow rapidly. Detection of Brain tumor is a serious issue in medical science. Imaging plays a central role in the diagnosis and treatment planning of brain tumor. The MR imaging method is the best due to its higher resolution. But there are many problems in detection of brain tumor in MR imaging as well. An important step in most medical imaging analysis systems is to extract the boundary of an area we are interested in. Many of the methods are there for the MRI segmentation [1-7].Though till now histogram thresholding is used for preprocessing only in many of the segmentation methods this paper shows that it can be used as a powerful tool for segmentation [2]. The image captured from a tumors brain shows the place of the infected portion of the brain. The image does not give the information about the numerical parameters such as area and volume of the infected portion of the brain. After preprocessing of the image, first image segmentation is done by using region growing segmentation. The segmented image shows the unhealthy portion clearly. From this image the infected portion (tumor) is selected by cropping the segmented image. From this cropped image, area is calculated [1]. II. RESEARCH WORK The authors in [2] compared the performance of classical sequential methods, a floating search method, and the “globally optimal” branch and bound algorithm when applied to functional MRI and intracranial EEG to classify pathological events. This work suggested that the sequential floating forward technique outperforms the other methodologies for these particular data. In terms of classification accuracy, the SFFS algorithm proved to be the best option for the automatic selection of features. Luts [3] proposed a technique to create Nosologic with help of color coding scheme for each voxel to distinguish distinctive tissues in a single image. For this purpose, a brain atlas and an abnormal tissue prior is acquired from MRSI data for segmentation. The detected abnormal tissue is then classified further by employing a supervised pattern recognition method followed by calculating the class probabilities for diverse tissue types. The proposed technique offers a novel way to visualize tumor heterogeneity in a specific image. The study results point out that combining MRI with MRSI feature improves classifiers’ performance. A prior for the abnormal tissue along with a healthy brain atlas further improves the nosologic images. Despite its usefulness, the proposed methodology, however, only provides the one-dimensional image features. Shi1et al. [4] employed neural networks for medical image processing, including the key features of medical image preprocessing, segmentation, and object detection and recognition. The study employed Hopfield and feed-forward neural networks. The feed-forward and Hopfield neural networks are simplest. The advantage of Hopfield neural networks is that it does not require pre-experimental knowledge. The time required to resolve image processing predicament is substantially reduced by using trained neural network. Kovacevic et al. [4] proposed a segmentation method for brain images that performs a basic segmentation process comprising three steps. In the first step, prominent features of images are extracted and normalization is carried out. In the next step, pixels are classified using artificial neural RESEARCH ARTICLE OPEN ACCESS