Federated Learning Surgery Dataset (FL_Surg)

This is the official benchmark dataset for the paper: Spatio-Temporal Representation Decoupling and Enhancement for Federated Instrument Segmentation in Surgical Videos.

It contains surgical video frames and segmentation masks collected from multiple clinical sites (simulated as site1 to site5), specifically designed for Federated Learning (FL) experiments in medical image segmentation. The dataset captures the heterogeneity of surgical instruments and environments across different centers.

Dataset Structure

The dataset is organized by Site -> Split -> Type (image/mask).

Directory Tree

FL_Dataset/
├── site1/
│   ├── train/
│   │   ├── image/ (Images *.png)
│   │   └── mask/  (Segmentation Masks *.png)
│   └── test/
│       ├── image/
│       └── mask/
├── site2/
│   ├── train/ ...
│   └── test/ ...
├── site3/ ...
├── site4/ ...
├── site5/ ...
└── metadata.csv  (Index file)