# Hard-Region Supervision: #1 on the Waymo Open Dataset 2D Video Panoptic Segmentation Leaderboard

arXiv: https://arxiv.org/abs/2609.38714 · HTML: https://mlsanity.com/paper/2609.38714/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 37 of its 44 references that are in the map.
- First posted on arXiv 30 Sep 2026.

## Builds on (most cited first, 30 of 37)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Mask R-CNN](https://mlsanity.com/paper/1703.06870/) (2017)
- [Masked-attention Mask Transformer for Universal Image Segmentation](https://mlsanity.com/paper/2112.01527/) (2021)
- [Cascade R-CNN: Delving into High Quality Object Detection](https://mlsanity.com/paper/1712.00726/) (2017)
- [Class-Balanced Loss Based on Effective Number of Samples](https://mlsanity.com/paper/1901.05555/) (2019)
- [LVIS: A Dataset for Large Vocabulary Instance Segmentation](https://mlsanity.com/paper/1908.03195/) (2019)
- [Per-Pixel Classification is Not All You Need for Semantic Segmentation](https://mlsanity.com/paper/2107.06278/) (2021)
- [Decoupling Representation and Classifier for Long-Tailed Recognition](https://mlsanity.com/paper/1910.09217/) (2019)
- [Long-tail learning via logit adjustment](https://mlsanity.com/paper/2007.07314/) (2020)
- [Simple Copy-Paste is a Strong Data Augmentation Method for Instance Segmentation](https://mlsanity.com/paper/2012.07177/) (2020)
- [Vision Transformer Adapter for Dense Predictions](https://mlsanity.com/paper/2205.08534/) (2022)
- [Born Again Neural Networks](https://mlsanity.com/paper/1805.04770/) (2018)
- [PointRend: Image Segmentation as Rendering](https://mlsanity.com/paper/1912.08193/) (2019)
- [Video Instance Segmentation](https://mlsanity.com/paper/1905.04804/) (2019)
- [Towards Understanding Ensemble, Knowledge Distillation and Self-Distillation in Deep Learning](https://mlsanity.com/paper/2012.09816/) (2020)
- [Equalization Loss for Long-Tailed Object Recognition](https://mlsanity.com/paper/2003.05176/) (2020)
- [Boundary loss for highly unbalanced segmentation](https://mlsanity.com/paper/1812.07032/) (2018)
- [Boundary IoU: Improving Object-Centric Image Segmentation Evaluation](https://mlsanity.com/paper/2103.16562/) (2021)
- [Ensemble Distribution Distillation](https://mlsanity.com/paper/1905.00076/) (2019)
- [Occluded Video Instance Segmentation: A Benchmark](https://mlsanity.com/paper/2102.01558/) (2021)
- [Seesaw Loss for Long-Tailed Instance Segmentation](https://mlsanity.com/paper/2008.10032/) (2020)
- [Video Panoptic Segmentation](https://mlsanity.com/paper/2006.11339/) (2020)
- [Equalization Loss v2: A New Gradient Balance Approach for Long-tailed Object Detection](https://mlsanity.com/paper/2012.08548/) (2020)
- [MinVIS: A Minimal Video Instance Segmentation Framework without Video-based Training](https://mlsanity.com/paper/2208.02245/) (2022)
- [Video K-Net: A Simple, Strong, and Unified Baseline for Video Segmentation](https://mlsanity.com/paper/2204.04656/) (2022)
- [DVIS: Decoupled Video Instance Segmentation Framework](https://mlsanity.com/paper/2306.03413/) (2023)
- [OMG-Seg: Is One Model Good Enough For All Segmentation?](https://mlsanity.com/paper/2401.10229/) (2024)

## Abstract (arXiv)

We describe our winning entry to the Waymo Open Dataset 2D Video Panoptic Segmentation Challenge. The task asks for a semantic class at every pixel of every frame and, for countable objects, an identity that holds across 100 frames and across five overlapping cameras. We build on DVIS++, a cascade of a segmenter, a tracker, and a refiner, as our baseline. We propose hard region supervision (HRS) to improve the baseline. In particular, we use the baseline to define the hard region as where it makes mistakes, and design a loss and an auxiliary prediction head for this region. The auxiliary head is used only in training and removed at test time, so at inference the model trained with HRS has the same architecture as the baseline. In addition, we propose three test-time steps that further improve the results: a two-model ensemble, a merge of the segmenter's output into the final panoptic map, and cross-camera identity linking. On the challenge test set, our entry reaches 0.3547 wSTQ, 0.2071 wAQ, and 0.6075 mIoU, ranking first on all three metrics. It is 3.6 wSTQ points ahead of the second entry and 2.4 points ahead of our DVIS++ baseline.