Research map / 3D vision / 3D generation
CVT-Occ: Cost Volume Temporal Fusion for 3D Occupancy Prediction
Zhangchen Ye, Tao Jiang, Chenfeng Xu, Yiming Li, Hang Zhao
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add temporal fusion (any of attention, warp-concat, plane-sweep) versus no temporal attention (ablation) | BEVFormer · Occ3D-Waymo val | mIoU: 23.87 -> 24.58 | Improved |
| replace temporal self-attention with cost volume temporal module (CVT-Occ) | BEVFormer · Occ3D-Waymo val | mIoU: 24.58 -> 27.37 | Improved |
| use CVT-Occ instead of warp-and-concat fusion | BEVFormer-WarpConcat · Occ3D-Waymo val | mIoU: 25.07 -> 27.37 | Improved |
| use CVT-Occ instead of SOLOFusion on road class | SOLOFusion · Occ3D-Waymo val Road | IoU: 70.95 -> 79.12 | Improved |
| increase total number of historical frames or frame interval (longer time span) | CVT-Occ · Occ3D-Waymo val | mIoU | Improved |
| increase temporal queue length | BEVFormer · Occ3D-Waymo val | mIoU | No clear effect |
| add CVT loss supervision on CVT module weights | CVT-Occ · Occ3D-Waymo val | mIoU | Improved |
| evaluate CVT-Occ in near region versus far region relative to BEVFormer (ablation) | CVT-Occ · Occ3D-Waymo val x range | mIoU: +3.76 near vs +1.43 far | Improved |
| evaluate CVT-Occ under fast versus slow ego speed relative to BEVFormer (ablation) | CVT-Occ · Occ3D-Waymo val speed split | mIoU: +3.17 fast vs +2.57 slow | Improved |
Later papers that compared with this method
1 later accepted paper reports CVT-Occ as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- FDR-Occ: Factorized Dense Routing for Full-Spectrum 3D Occupancy Prediction (ECCV 2026 · better on 1 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- ViPOcc: Leveraging Visual Priors from Vision Foundation Models for Single-View 3D Occupancy Prediction (AAAI 2025 · similarity 0.90)
- RadOcc: Learning Cross-Modality Occupancy Knowledge through Rendering Assisted Distillation (AAAI 2024 · similarity 0.89)
- Spatiotemporal Decoupling for Efficient Vision-Based Occupancy Forecasting (CVPR 2025 · similarity 0.89)
- Rethinking Temporal Fusion with a Unified Gradient Descent View for 3D Semantic Occupancy Prediction (CVPR 2025 · similarity 0.88)
- OccProphet: Pushing the Efficiency Frontier of Camera-Only 4D Occupancy Forecasting with an Observer-Forecaster-Refiner Framework (ICLR 2025 · similarity 0.87)
- SDGOCC: Semantic and Depth-Guided Bird's-Eye View Transformation for 3D Multimodal Occupancy Prediction (CVPR 2025 · similarity 0.87)
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.
