Research map / Video / Video segmentation and tracking
4D Spatio-Temporal ConvNets: Minkowski Convolutional Neural Networks
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 4D MinkowskiNet instead of 3D MinkowskiNet | MinkowskiNet20 · Synthia 4D | score 1: 76.24 -> 77.46 | Helped |
| use 4D MinkowskiNet instead of 3D MinkowskiNet | MinkowskiNet20 · Synthia 4D | score 2: 89.31 -> 88.013 | Hurt |
| add TS-CRF to 4D MinkNet20 (ablation) | 4D MinkNet20 · Synthia 4D | score 1: 77.46 -> 78.30 | Helped |
| add TS-CRF to 4D MinkNet20 (ablation) | 4D MinkNet20 · Synthia 4D | score 2: 88.013 -> 90.23 | Helped |
| replace the hybrid kernel with a tesseract kernel (ablation) | 4D MinkowskiNet20 · Synthia 4D | score 1: 77.46 -> 75.34 | Hurt |
| replace the hybrid kernel with a tesseract kernel (ablation) | 4D MinkowskiNet20 · Synthia 4D | score 2: 88.013 -> 89.27 | Helped |
| add temporal aggregation (TA) to 3D MinkNet20 (ablation) | 3D MinkNet20 · Synthia 4D | score 1: 76.24 -> 77.03 | Helped |
| add temporal aggregation (TA) to 3D MinkNet20 (ablation) | 3D MinkNet20 · Synthia 4D | score 2: 89.31 -> 89.20 | Hurt |
| use 4D MinkowskiNet instead of 3D MinkowskiNet | MinkowskiNet42 · noisy Synthia 4D | overall score: 76.717 -> 79.836 | Helped |
| increase network depth from MinkNet20 to MinkNet32 and use TS-CRF (ablation) | 4D MinkowskiNet + TS-CRF · Synthia 4D | score 1: 78.30 -> 78.67 | Helped |
| use 4D MinkowskiNet instead of 3D MinkowskiNet | MinkowskiNet · RueMonge 2014 TASK3 | score: 66.46 -> 66.56 | No clear effect |
| add TS-CRF to 4D MinkNet20 (ablation) | 4D MinkNet20 · RueMonge 2014 TASK3 | score: 66.56 -> 66.59 | No clear effect |
Closest papers by meaning
- V4D: 4D Convolutional Neural Networks for Video-level Representation Learning (ICLR 2020 · similarity 0.89)
- 3D Semantic Segmentation With Submanifold Sparse Convolutional Networks (CVPR 2018 · similarity 0.89)
- Fused View-Time Attention and Feedforward Reconstruction for 4D Scene Generation (NeurIPS 2025 · similarity 0.86)
- LSK3DNet: Towards Effective and Efficient 3D Perception with Large Sparse Kernels (CVPR 2024 · similarity 0.86)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.86)
- OA-CNNs: Omni-Adaptive Sparse CNNs for 3D Semantic Segmentation (CVPR 2024 · similarity 0.86)
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.