Presentation description

Medical image segmentation requires inputting multiple user prompts, such as points, bounding boxes, and scribbles. When segmenting such medical images, this can often times be very time consuming as the amount of needed user input increases. Knowledge distillation helps in this, which is a technique that transfers knowledge from a larger, more complex model (the teacher) to a smaller, more efficient model. This in turn reduces the amount of user input burden required for accurate segmentation.