Arthrex
AutoScribe

Helping Surgeons Help You

Project Type: 
Electrical Engineering
Year: 
2019

Description

Arthroscopic surgery is a minimally invasive procedure, increasingly preferred by medical professionals and patients due to its advantages over traditional open surgery. Arthroscopic operations are performed 2 million times per year worldwide, and 750,000 times per year in the U.S. alone. During arthroscopy, the surgeon makes 3 small incisions into a joint, inserting surgical tools and a tiny camera. The camera displays the inside of the joint on an external monitor as well as records the procedure, which is later annotated by surgeons for diagnosis, patient debriefing, and training future medical professionals. By utilizing machine learning algorithms, we can automate post-operative annotation,  allowing surgeons to focus more time on surgery and less time tediously captioning videos. Arthrex AutoScribe is a machine learning algorithm that detects which tool is being used and labels the video. By labeling the video with the specific tool, surgeons can then easily determine what event is happening in the procedure for more descriptive captioning from the relevant video segments.

Automatically labeling video is only a small part in the world of arthroscopic surgery. We believe creating a high accuracy algorithm paves the way for automated video captioning, alleviating orthopaedic surgeons’ postoperative responsibilities, and enabling them to spend the time necessary to more effectively treat their patients.

Students

Electrical Engineering
Electrical Engineering
Electrical Engineering