NECEC 2026 · Poster

Benchmarking Deep-Learning and Foundation-Model Segmentation for UAV Kelp Canopy Mapping on the Newfoundland Coast

Syed Omar Faruk, Zintis Stasko, Patrick Gagnon, Lourdes Peña-Castillo, Christopher Cogan · Memorial University of Newfoundland

Links

Pretrained-model viewerBrowse the stored model outputs tile by tile on the drone mosaics of four sites, compare models and thresholds, and inspect pixels. Open for the conference period only.

The SAV Finder tool (run the models on a tile) is not public. To request access, email sofaruk@mun.ca.

References

  1. Breiman, L. (2001). Random forests. Machine Learning 45, 5–32. doi:10.1023/a:1010933404324
  2. Ronneberger, O., Fischer, P. & Brox, T. (2015). U-Net: convolutional networks for biomedical image segmentation. MICCAI. doi:10.1007/978-3-319-24574-4_28
  3. Chen, L.-C., Papandreou, G., Schroff, F. & Adam, H. (2017). Rethinking atrous convolution for semantic image segmentation. arXiv:1706.05587. doi:10.48550/arXiv.1706.05587
  4. Jocher, G., Chaurasia, A. & Qiu, J. (2023). Ultralytics YOLOv8 [software]. github.com/ultralytics/ultralytics
  5. Oquab, M., et al. (2023). DINOv2: learning robust visual features without supervision. arXiv:2304.07193. doi:10.48550/arXiv.2304.07193
  6. Ravi, N., et al. (2024). SAM 2: Segment anything in images and videos. arXiv:2408.00714. doi:10.48550/arXiv.2408.00714
  7. Brodersen, K. H., Ong, C. S., Stephan, K. E. & Buhmann, J. M. (2010). The balanced accuracy and its posterior distribution. ICPR. doi:10.1109/icpr.2010.764

Contact

Syed Omar Faruk · sofaruk@mun.ca