Research map / 3D vision / Face and shape
Learning Attentive and Hierarchical Representations for 3D Shape Recognition
Jiaxin Chen, Jie Qin, Yuming Shen, Li Liu, Fan Zhu, Ling Shao
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 view-agnostic attention (VAA) (ablation) | Baseline MVCNN · ModelNet40 retrieval (VGG-A) | mAP: 72.9 -> 86.3 | Improved |
| add view-specific attention (VSA) (ablation) | Baseline MVCNN · ModelNet40 retrieval (VGG-A) | mAP: 72.9 -> 87.5 | Improved |
| add hybrid attention (VAA+VSA) (ablation) | Baseline MVCNN · ModelNet40 retrieval (VGG-A) | mAP: 72.9 -> 89.7 | Improved |
| add multi-granular view pooling (MVP) (ablation) | Baseline MVCNN+HA · ModelNet40 retrieval (VGG-A) | mAP: 89.7 -> 90.0 | Improved |
| add hyperbolic embedding and hyperbolic MLR (ablation) | Baseline MVCNN+HA+MVP · ModelNet40 retrieval (VGG-A) | mAP: 90.0 -> 91.1 | Improved |
| add view-agnostic attention (VAA) (ablation) | Baseline MVCNN · ModelNet10 retrieval (VGG-A) | mAP: 80.1 -> 88.7 | Improved |
| add view-specific attention (VSA) (ablation) | Baseline MVCNN · ModelNet10 retrieval (VGG-A) | mAP: 80.1 -> 90.2 | Improved |
| add hybrid attention (VAA+VSA) (ablation) | Baseline MVCNN · ModelNet10 retrieval (VGG-A) | mAP: 80.1 -> 92.8 | Improved |
| add multi-granular view pooling (MVP) (ablation) | Baseline MVCNN+HA · ModelNet10 retrieval (VGG-A) | mAP: 92.8 -> 93.0 | Improved |
| add hyperbolic embedding and hyperbolic MLR (ablation) | Baseline MVCNN+HA+MVP · ModelNet10 retrieval (VGG-A) | mAP: 93.0 -> 94.2 | Improved |
| switch backbone from VGG-M to ResNet-50 | HEAR · ModelNet40 classification | per-instance accuracy: 95.5 -> 96.7 | Improved |
| switch backbone from VGG-A to ResNet-50 | HEAR · ModelNet40 retrieval | mAP: 91.1 -> 92.0 | Improved |
Closest papers by meaning
- Learning Canonical View Representation for 3D Shape Recognition With Arbitrary Views (ICCV 2021 · similarity 0.85)
- MVTN: Learning Multi-view Transformations for 3D Understanding (IJCV 2024 · similarity 0.85)
- MLVCNN: Multi-Loop-View Convolutional Neural Network for 3D Shape Retrieval (AAAI 2019 · similarity 0.85)
- Shape2Scene: 3D Scene Representation Learning Through Pre-training on Shape Data (ECCV 2024 · similarity 0.85)
- Emphasizing 3D Properties in Recurrent Multi-View Aggregation for 3D Shape Retrieval (AAAI 2018 · similarity 0.85)
- TAMM: TriAdapter Multi-Modal Learning for 3D Shape Understanding (CVPR 2024 · similarity 0.83)
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.
