DropLoss for Long-Tail Instance Segmentation


  • Ting-I Hsieh National Tsing Hua University
  • Esther Robb Virginia Tech
  • Hwann-Tzong Chen National Tsing Hua University Aeolus Robotics
  • Jia-Bin Huang Virginia Tech




Object Detection & Categorization


Long-tailed class distributions are prevalent among the practical applications of object detection and instance segmentation. Prior work in long-tail instance segmentation addresses the imbalance of losses between rare and frequent categories by reducing the penalty for a model incorrectly predicting a rare class label. We demonstrate that the rare categories are heavily suppressed by correct background predictions, which reduces the probability for all foreground categories with equal weight. Due to the relative infrequency of rare categories, this leads to an imbalance that biases towards predicting more frequent categories. Based on this insight, we develop DropLoss -- a novel adaptive loss to compensate for this imbalance without a trade-off between rare and frequent categories. With this loss, we show state-of-the-art mAP across rare, common, and frequent categories on the LVIS dataset. Codes are available at https://github.com/timy90022/DropLoss.




How to Cite

Hsieh, T.-I., Robb, E., Chen, H.-T., & Huang, J.-B. (2021). DropLoss for Long-Tail Instance Segmentation. Proceedings of the AAAI Conference on Artificial Intelligence, 35(2), 1549-1557. https://doi.org/10.1609/aaai.v35i2.16246



AAAI Technical Track on Computer Vision I