International Journal of Computer Engineering and Applications, Volume XII, Issue I, Jan. 18, www.ijcea.com ISSN 2321-3469 Sanjukta Mishra 31 A REVIEW ON BIG DATA ANALYTICS IN MEDICAL IMAGING Sanjukta Mishra Assistant Professor, Department of Computer Science, Brainware University,Kolkata, India ABSTRACT: Big data analytics in the application of medical image processing plays a vital role .For better healthcare it is necessary for automatic analysis which helps in early diagnosis of diseases. This paper describes the overview of Big Data, the benefits, challenges and models of big data analytics for analyzing the medical images from a large dataset in a distributed environment. This paper also proposed a Spark with Yarn framework for medical imaging. Keywords: Big Data, Hadoop, Spark, Yarn, Medical imaging [1] INTRODUCTION The medical industry has generated huge amounts of data that can be in hard copy or in soft copy form. The main focus is to store, manage and analyze the massive amounts of data. The rate of growing size of medical images is very high. The medical image datasets named Image CLEF contained approx 66,000 images between 2005 and 2007 which had increased to approx 300,000 images in 2013[1].The images are also in different varieties, modalities, dimensions, resolutions and qualities. So the new challenges arise like proper data extraction, data integration and data mining. More research needs to be done for the multimodal image analysis. The unstructured medical images need to be structured for accurate data mining and proper analysis. The medical images data can be collected from different institutions or organizations. So to utilize such types of data some important analytic models need to be developed to evaluate and validate .If proper data has been extracted and evaluated from the large database or big data the early diagnosis of diseases possible which reduces the risk of life and the medical cost. It is possible that for the same applications like Alzheimer and the same modality like PET scan ,various organizations may use different models in image acquisitions,