Robotic Scene Segmentation Sub-Challenge
Part of the Endoscopic Vision Challenge
In 2017 we ran the Robotic Instrument Segmentation Sub-Challenge and had 10 teams participate at the challenge day in Quebec City with some great results. Since then 85 individuals have downloaded our data and we hope to see some interesting publications come out of it in the near future.
This year we want to extend the challenge complexity to include the entire scene, where we hope to have users label robotic instruments as well as anatomical objects and non-robotic surgical instruments such as suturing needles and gauze. In our 2 upcoming data relases we will provide densely labelled surgical images from 8 different robotic procedures. At the end of the summer 2 unseen sequences will be released as the test set.