Deepwound: automated postoperative wound assessment and surgical site surveillance through convolutional neural networks

Varun Shenoy, California, USA 16-18

My research investigates the use of convolutional neural networks and a machine learning algorithm in image classification for surgical wound surveillance. I developed a computational model to accurately identify the presence of nine mutually inclusive labels in a wound image, including surgical site infection, drainage, staples, and sutures. Moreover, I built a mobile application that pairs with my classification models so patients can track their wound from the comfort of their home along with other important recovery factors, such as exercise and medicine intake. This information can be compiled into a document to be shared with the patient's health support team.

Show more