Research map / Detection and segmentation / Open-set and few-shot detection
Hierarchical Lovasz Embeddings for Proposal-Free Panoptic Segmentation
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use shared hierarchical Lovasz embeddings rather than associative embedding loss (ablation) | HLE · Cityscapes val | PQ: 38.8 -> 59.1 | Helped |
| Use shared hierarchical Lovasz embeddings rather than cross-entropy hierarchical embeddings (ablation) | HLE · Cityscapes val | PQ: 45.0 -> 59.1 | Helped |
| Keep semantic and instance features in one joint embedding space rather than split halves (ablation) | HLE · Cityscapes val | PQ: 50.2 -> 59.1 | Helped |
| Use a joint embedding instead of a separate semantic segmentation branch (ablation) | HLE · Cityscapes val | PQ: 57.0 -> 59.1 | Helped |
| Add VQ-VAE style semantic-mean loss (ablation) | HLE · Cityscapes val | PQ: 57.0 -> 59.1 | Helped |
| Add Thomson initialization for semantic means (ablation) | HLE · Cityscapes val | PQ: 58.6 -> 59.1 | Helped |
| Downsample the embedding space during post-processing | HLE · Cityscapes val | PQ: 59.8 -> 58.4 | Hurt |
| Use HLE with ResNet50 instead of DeeperLab with Xception71 | Cityscapes val · Cityscapes val | PQ: 56.5 -> 59.8 | Helped |
Closest papers by meaning
- Unifying Training and Inference for Panoptic Segmentation (CVPR 2020 · similarity 0.89)
- Panoptic, Instance and Semantic Relations: A Relational Context Encoder To Enhance Panoptic Segmentation (CVPR 2022 · similarity 0.89)
- LPSNet: A Lightweight Solution for Fast Panoptic Segmentation (CVPR 2021 · similarity 0.89)
- Video Panoptic Segmentation (CVPR 2020 · similarity 0.88)
- Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation (CVPR 2020 · similarity 0.88)
- Scene-Centric Unsupervised Panoptic Segmentation (CVPR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.