Declining medical imaging Innovations and a greater dependence on high-resolution images and videos for diagnosis have resulted during a medical data explosion. Frost & Sullivan estimates that by 2020, the amount of information generated and archived within the us alone will exceed 2,500 petabytes (1 petabyte equals 1 million gigabytes). This necessitates a radically different approach to handling data, which conventionally relies on monolithic data archives. a giant Data-based medical imaging stack would come with both structured and unstructured images, requiring interaction with a spread of vendors and solution providers that handle various aspects of information capture, storage, management, analysis, reporting and decision support. The market is dominated by data companies IBM, McKesson, Lexmark and Dell, and healthcare companies like GE, Philips and Siemens. the following logical avenue for innovation would be to form use of the petabytes of knowledge being generated daily. to create sense of the vast and diverse data types—images, text, video, and reference databases—is no mean (or manual) feat. this can be where Big Data analytics platforms like Hadoop and NoSQL, and AI tools inherit play. rather than subjecting diagnosticians to a deluge of data—and conversely subjecting complex data to a poor laboratory technician—advanced algorithms can help within the analysis and interpretation of medical images. Computer-aided diagnosis is already popular in clinical practices; this might evolve into a computer-only diagnosis. it's possible that machine learning algorithms will take over the task of interpreting images, comparing test results with reference standards and therefore the patient’s own past records, and presenting radiologists with outcomes. In essence, radiologists will cease viewing medical imaging, basing their clinical decisions solely on information supplied by advanced algorithms.