Brain Tumor Classification

DETECTION AND CLASSIFICATION OF BRAIN TUMOR IMAGES USING BACK PROPAGATION FUZZY NEURAL NETWORK

N. Periyasamy1, Dr. J. G. R. Sathiaseelan 2

1Research Scholar, Department of Computer Science, Bishop Heber College, Tiruchirappalli - 620017

2 Head, Department of Computer Science, Bishop Heber College, Tiruchirappalli – 6200177

[e-mail]

[e-mail]

Abstract

Artificial Neural networks is a substantial research area in medical image classification. The Bio Medical image recognition technique have been generally applied in various diagnosis diseases to predict the result most accurate result. This paper illustrates the structure of the maintenance for the image classification process stages of a brain tumor as

There are many different types of brain tumor, which make the decision very complicated. Hence the classification of brain tumor images is very essential to classify the brain tumor, which affects the lives of the people. The classification of tumor process can lead to the right decision and provide correct and appropriate treatment. Treatment of the brain tumor mostly depends on its size and its types. Treatment can be different for the two types of the tumor, which mostly depends on the factors like Age, overall health, medical history, Tumor type, Location of the tumor and its size. The memory requirement for the brain tumor dataset is much less than MRI images. Produced dataset is endangered to several clustering and classification approaches [1]. Many canvassers have recommended dissimilar techniques for the classification of brain tumors based on discrete wavelet transform, Bayesian neural network, K-Nearest Neighbor, PCA and probabilistic neural network. The comprehensive features and the fuzzy rules are implemented to classify an abnormal brain image to the corresponding tumor [2]. In this paper, we propose an approach for building a classification model for brain tumors. Regions of interest (tumor) are first segmented. Then, it is based on the abnormality of the tumor classification and segmentation are done. In this paper, we took the training samples from other sources. This leads to improved classification accuracy. In brain MR images, after appropriate segmentation of brain tumor classification of tumor into malignant and benign is difficult task due to the complexity and variations in tumor tissue characteristics like its gray level of the image, the shape of the image and size of the image