Tool Detection & Segmentation In Surgical Room via YOLO5

This repo contains an easy-to-use model for detecting & segmenting tools in both hans of a suturing simulator.

Getting Started

Dependencies

• The dependecies are list via the ruquirments.txt file in the repo.

Executing program

• Segmenting a video
  • Your video file should be located in a "videos" directory

python video.py --video_file_name videos\[your_video_name] --model_path models\[your_model_name]

• Segmenting a video with evaluation
  • Your video file should be located in a "videos" directory
  • Your label direrctory should consists two sub-directories: tools_right & tools_left

python video.py --video_file_name videos\[your_video_name] --model_path models\[your_model_name] --video_labels_path [path_to_label_dir]

• Segmenting an image
  • Your image file should be located in a "images" directory

python predict.py --image_file_name images\[your_video_name] --model_path models\[your_model_name]

Authors

• @Ido Levi
• @Daniel Yehezkel

Version History

• 0.1
  • Initial Release