Home

Awesome

blob loss: instance imbalance aware loss functions for semantic segmentation

<img width="1408" alt="Screenshot 2023-02-14 at 00 05 30" src="https://user-images.githubusercontent.com/7048826/218594471-04e4e64a-647c-47f4-9927-0f4936c61987.png">

example implementation - computation time

note that this example implementation is not optimized for computation time. We would love to see your more efficient implementation!

manuscript

https://arxiv.org/abs/2205.08209

citation

Please cite blob loss when using it:

@misc{https://doi.org/10.48550/arxiv.2205.08209,
  doi = {10.48550/ARXIV.2205.08209},
  
  url = {https://arxiv.org/abs/2205.08209},
  
  author = {Kofler, Florian and Shit, Suprosanna and Ezhov, Ivan and Fidon, Lucas and Horvath, Izabela and Al-Maskari, Rami and Li, Hongwei and Bhatia, Harsharan and Loehr, Timo and Piraud, Marie and Erturk, Ali and Kirschke, Jan and Peeken, Jan and Vercauteren, Tom and Zimmer, Claus and Wiestler, Benedikt and Menze, Bjoern},
  
  keywords = {Computer Vision and Pattern Recognition (cs.CV), Machine Learning (cs.LG), Image and Video Processing (eess.IV), FOS: Computer and information sciences, FOS: Computer and information sciences, FOS: Electrical engineering, electronic engineering, information engineering, FOS: Electrical engineering, electronic engineering, information engineering},
  
  title = {blob loss: instance imbalance aware loss functions for semantic segmentation},
  
  publisher = {arXiv},
  
  year = {2022},
  
  copyright = {arXiv.org perpetual, non-exclusive license}
}

evaluation

HINT: Our new project panoptica might be helpful in evaluating your multi-instance segmentation problem.