Two Dimensional Medical Images Diagnosis using MapReduce

Objectives: Due to advanced camera capturing techniques used in medical domain efficient management and quick diagnosis of massively generated 2D/3D medical data has become challenging tasks for doctors. Methods: In this paper, we propose an idea for analyzing medical images using Hadoop's MapReduce Framework. HDFS is used for storing feature library of existing medical images and parallelism in indexing; matching and retrieval processes are achieved by MapReduce. The Map function is used to match feature vector of the query image with feature vectors present in feature library, while the Reduce function is used to aggregate and sort results from all the mappers. Findings: As a result of parallelism Hadoop based medical image retrieval system will take very less time for image retrieval as compared to the traditional image retrieval systems. Existing Content based medical image retrieval system uses only image processing methods, but to handle Large scale image retrieval and indexing processes we are using MapReduce. It results in the elimination of manual processes of diagnosis and leads to automated detection and diagnosis. Application: This model will definitely help doctors in real time decision making and understanding of particular stage of disease. Also real time treatment is suggested after identification of a particular type of disease.