Abstract: It hasalways been the ardent pursuit of man to extend his capabilities through toolsand technology, and one of the most promising technologies developed in therecent years is Computer Vision (CV). Though we have come a long way in this technology, image segmentation orscene understanding still remains one of the pivotal, yet a very difficult taskin the field of CV. Lately, theavailability of powerful graphics processing units coupled with complex imageprocessing algorithms made CV a viable and promising tool. With the advent of deep neural networks,pixel-level segmentation and object recognition are reaching human-level speedsand accuracies, if not better. Thebenefits thereof, especially in the medical field, are only limited by ourimagination. Its prospects are beingexplored in every phase of the medical field, like in early identification ofdiseases, faster and better diagnostics, impeccable micro-surgical abilities,better prediction of disease recurrence or patient recovery time, and reliablemanagement suggestions in almost every specialty. In this paper, we attempt to review the recentwork on semantic segmentation using deep neural networks in the field ofmedicine. Semantic segmentation has beenstudied on a wide range of anatomical systems, especially since well-trainedmodels like ImageNet have been available since 2012. In just the last 5 to 10 years, there havebeen at least 30 well researched articles published in this field, some ofwhich are very promising with highest accuracies achieved in their respectivefields. As expected, the most explored fieldseems to be cancer, in various medical systems like neurology, respiratory,reproductive, etc., while other systems generally have also been studied likecardiovascular, endocrine, urinary, etc. Other interesting applications were in pretreatment risk analysis andforensic medicine.
Keywords: MedicalImaging, Computer Vision, Deep Learning, Machine Learning, Convolutional NeuralNetworks, Semantic Segmentation