Skip to content

Latest commit

 

History

History
31 lines (21 loc) · 1.24 KB

README.md

File metadata and controls

31 lines (21 loc) · 1.24 KB
Raw

Activity Phase Detection in Cholecystectomy Surgery Videos [Done as part of CS766 course at UW-Madison]

For more information and results, please check the project website - Website

This codebase is built upon the code present here - https://github.com/JayJayBinks1/DeepPhase

So, I would like to acknowledge and credit the user above for the code.

Instructions to run the code

  1. Download the dataset from http://camma.u-strasbg.fr/datasets
  2. Put all videos in a yourpath/cholec80/Data folder, where yourpath is a local path on your machine. This can be changed in config.py file with DATA_DIR value.

For example, all videos will have path like - yourpath/cholec80/Data/video01.mp4

Tool annotations will have path like - yourpath/cholec80/Data/video01-tool.txt

Phase annotations will have path like - yourpath/cholec80/Data/video01-phase.txt

Timestamps will have path like - yourpath/cholec80/Data/video01-timestamp.txt

  1. Run frame_extraction.py to generate frames from videos.
  2. Run tool_recognition.py for tool recognition.
  3. Run feature_extract.py for feature extraction.
  4. Run phase_recognition.py for phase recognition.