Corrigendum to “A novel multi-attention, multi-scale 3D deep network for coronary artery segmentation”(Medical Image Analysis, (2023), 85, C, (102745), (S1361841523000063), 10.1016/j.media.2023.102745)

The authors regret that the public code repository (GitHub: Cassie-CV/CAS-Net) associated with this article was incomplete at the time of publication. The initial release contained only baseline models used for comparison (e.g., UNet and CS²-Net) while omitting the core implementation of the proposed CAS-Net (specifically, the AGFF, SAFE, and MSFA modules). The authors also regret that several general-purpose utility scripts (e.g., evaluation_metrics3D.py) reused from the publicly available CS²-Net codebase were not explicitly acknowledged in the original release. The repository has now been fully updated to include:1.The complete and verified CAS-Net implementation, consistent with the methodology described in the published article.2.A detailed README.md file providing environment setup, training/testing instructions, and explicit acknowledgments for all reused components. The complete and verified CAS-Net implementation, consistent with the methodology described in the published article. A detailed README.md file providing environment setup, training/testing instructions, and explicit acknowledgments for all reused components. The corrected repository is publicly available at: https://github.com/Cassie-CV/CAS-Net These corrections do not affect any results, figures, or conclusions presented in the published paper. The authors would like to apologise for any inconvenience caused.