Research map / Detection and segmentation / 3D detection and driving
VIP-DeepLab: Learning Visual Perception With Depth-Aware Video 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 ViP-DeepLab instead of VPSNet | Cityscapes-VPS val · Cityscapes-VPS val | VPQ: 57.5 -> 63.1 | Helped |
| use ViP-DeepLab instead of VPSNet | Cityscapes-VPS test · Cityscapes-VPS test | VPQ: 57.4 -> 62.5 | Helped |
| use ViP-DeepLab instead of MPSD | KITTI depth leaderboard · KITTI depth leaderboard | SILog: 11.12 -> 10.80 | Helped |
| use ViP-DeepLab instead of PointTrack | KITTI MOTS leaderboard pedestrians · KITTI MOTS leaderboard pedestrians | sMOTSA: 61.5 -> 67.7 | Helped |
| use ViP-DeepLab instead of PointTrack | KITTI MOTS leaderboard cars · KITTI MOTS leaderboard cars | sMOTSA: 78.5 -> 80.6 | Helped |
| initialize from Mapillary Vistas-pretrained checkpoint (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 58.5 -> 59.0 | Helped |
| further pretrain on Cityscapes videos with pseudo labels (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 59.0 -> 59.9 | Helped |
| increase next-frame instance branch context modules from 1 to 4 (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 59.9 -> 60.9 | Helped |
| apply AutoAugment (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 60.9 -> 61.3 | Helped |
| add Recursive Feature Pyramid (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 61.3 -> 61.9 | Helped |
| use multi-scale inference and horizontal flipping (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 61.9 -> 62.7 | Helped |
| add temporally consistent pseudo-labeled training samples (ablation) | ViP-DeepLab · Cityscapes-VPS | VPQ: 62.7 -> 63.1 | Helped |
Closest papers by meaning
- PanopticDepth: A Unified Framework for Depth-Aware Panoptic Segmentation (CVPR 2022 · similarity 0.91)
- Panoptic-DeepLab: A Simple, Strong, and Fast Baseline for Bottom-Up Panoptic Segmentation (CVPR 2020 · similarity 0.89)
- Scene-Centric Unsupervised Panoptic Segmentation (CVPR 2025 · similarity 0.86)
- PanoRecon: Real-Time Panoptic 3D Reconstruction from Monocular Video (CVPR 2024 · similarity 0.85)
- PanoOcc: Unified Occupancy Representation for Camera-based 3D Panoptic Segmentation (CVPR 2024 · similarity 0.85)
- IPFormer: Visual 3D Panoptic Scene Completion with Context-Adaptive Instance Proposals (NeurIPS 2025 · similarity 0.85)
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.