AN EFFECTIVE USER INTERFACE IMAGE PROCESSING MODEL FOR CLASSIFICATION OF BRAIN MRI TO PROVIDE PROLIFIC HEALTHCARE

Author Name(s): Vinit Kumar Gunjan, M Janga Reddy, Fahimuddin.Shaik, V. Hymavathi
Author Email: vinitkumargunjan@gmail.com

Abstract

Today the design and development of computer assisted User Interface Image processing Systems have been very helpful in detecting and diagnosing the underlying diseases in the early stage. Out of many anomalous diseases, Brain Tumor Diagnosis and treatment has been a challenge to medical as well as to research community. If it is detected at an early stage the chances of saving a life are more. The primary goal of this work is to develop an effective User Interface Image Processing Model via a GUI (Graphical User Interface) for brain tumor detection and segmentation. The methods amalgamated together and investigated here are Fast Bounding Box (FBB) method and Discrete Wavelet Transform (DWT) with Morphology & SVM (Support Vector Machine) Classifiers.

INTRODUCTION

The uncharacteristic growth of cells in the brain prompts Brain tumor. Normally brain tumor rises up from brain cells, blood vessels or nerves that are existing in the brain. Early recognition of brain tumor is vital as death rate is higher among people having brain tumor [1]. As per 2007 estimates absolutely 80,271 people are affected by different sorts of tumor in India [6].Techniques for brain tumor identification utilizing image processing have been available for couple of decades. Researchers have proposed numerous semi-automatic in addition to automatic image processing techniques for identifying brain tumors however the majority of them fail to give effective and exact outcomes because of the presence of noise, inhomogeneity, poor images contrast that arise typically in medical images [4].                                            The significance of this proposal is to design and develop a complete automated diagnosis system for brain tumor, including removal of noise, detection of the brain tumor, extraction as well as classification of brain tumor. This paper proposes amalgamation of two methods, Fast Bounding Box method and DWT with Morphology & SVM Classifiers . Even though some studies suggest that segmentation of brain tumor from MRI Images is time consuming procedure but it is an important research to be carried out [5].

Conclusion

In this work, a coronal fat-suppressed T1-weighted magnetic resonance image is considered for simulation, the coronal image which is considered for this work is obtained after gadolinium administration. The outcomes got have been analyzed with respect to the ROI (damaged part of the brain) which has been observed. This research work is primarily useful as an automatic CAD system for brain Tumor segmentation. As a future scope the images resulted can be verified for statistical approaches to find image attributes which will be helpful in parametric analysis.

1,185 total views, no views today

Download PDF File

About the author: admin