Of an order of 330 million surgeries are performed worldwide every year. Yet there is a backlog of around 150 million pending surgeries annually. Surgical robotics is becoming a lot more sophisticated by the day. Much of this increasing success is due to advances in Computer Vision (CV) . CV allows the tracking of tools, detection of organs, and a description of the phase of surgery that enables a surgeon to perform the delicate art of surgery with much greater precision and efficiency. These advancements have also enabled remote robotic surgery in which a patient and the surgeon can be far apart from each other geographically.

In this paper, we report promising results of evolving the architecture of capsule networks for tool classification for Computer-Aided Laparoscopy (CAL).