Three-Dimensional MRI Brain Image Analysis on Hadoop Platform

摘要

Tremendous medical data generation takes place in hospitals due to advanced image capturing techniques. Real-time analysis of that large knowledge becomes essential for medical practitioners. This paper focuses on implementation of automatic diagnosis of three-dimensional data using the MapReduce framework. After flourishing completion of two-dimensional image analysis of magnetic resonance imaging (MRI) brain pictures for detecting tumor, we have introduced 3D image analysis of brain MRI for tumor detection. Every image in diagnosis is analyzed by well-known Scale Invariant Feature Transform (SIFT) rule to spot keypoints at varied scales, whereas Hadoop's MapReduce is employed to match extracted feature vector of every inputted image, and HDFS supports storing options of existing medical pictures within the system. This application will certainly cut back diagnosis time and facilitate doctors for higher decision making.