Dataset of placental vessel segmentation for Twin-to-Twin Transfusion Syndrome

This is the dataset related to the paper:

Płotka, Szymon, et al. "Real-time placental vessel segmentation in fetoscopic laser surgery for Twin-to-Twin Transfusion Syndrome." Medical Image Analysis 99 (2025): 103330.

We release the TTTSNet dataset, a fetoscopic video dataset consisting of 4,408 frames (2,060 for training and validation, and 2,348 for testing), obtained from 42 independent patients who underwent TTTS fetoscopic procedures at six European fetal surgery centers. Data from centers A and B are publicly available, while data from centers C, D, E, and F are newly released to the community to foster research in this domain.

Please note: The FetReg2021 dataset (from Centers A and B), which is currently publicly accessible, contains pixel-wise annotations that omit small placental vessel segmentation and include incomplete labels for larger vessels. For the purposes of this paper, these segmentations were revised to align with our segmentation protocol.

If you use this dataset, please cite our research:

@article{plotka2025real,
  title={Real-time placental vessel segmentation in fetoscopic laser surgery for Twin-to-Twin Transfusion Syndrome},
  author={P{\l}otka, Szymon and Szczepa{\'n}ski, Tomasz and Szenejko, Paula and Korzeniowski, Przemys{\l}aw and Calvo, Jes{\'u}s Rodriguez and Khalil, Asma and Shamshirsaz, Alireza and Brawura-Biskupski-Samaha, Robert and I{\v{s}}gum, Ivana and S{\'a}nchez, Clara I and others},
  journal={Medical Image Analysis},
  volume={99},
  pages={103330},
  year={2025},
  publisher={Elsevier}
}