Research map / 3D vision / Face and shape
Angular Triplet-Center Loss for Multi-View 3D Shape Retrieval
Read the paperCheck 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 |
|---|---|---|---|
| Optimize embeddings with ATCL instead of TCL's Euclidean distance (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 84.35% -> 85.35% | Helped |
| Add ATCL to softmax supervision (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 78.28% -> 86.11% | Helped |
| Add softmax loss to ATCL (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 85.35% -> 86.11% | Helped |
| Replace softmax loss with softmax loss plus center loss (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 78.28% -> 80.95% | Helped |
| Use ATCL rather than softmax loss alone (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 78.28% -> 85.35% | Helped |
| Increase angular margin beyond 1 (ablation) | ATCL · ModelNet40 | MAP | Hurt |
| Set angular margin to 0.7 (ablation) | ATCL · ModelNet40 | AUC: 86.48% | Helped |
| Set angular margin to 0.7 (ablation) | ATCL · ModelNet40 | MAP: 85.21% | Helped |
| Increase ATCL loss weight from 0.01 toward the 1:1 setting (ablation) | ATCL+softmax · ModelNet40 | retrieval performance | Mixed |
| Use ATCL+softmax rather than GIFT | VGG-M retrieval model · ModelNet10 | MAP: 91.12% -> 92.07% | Helped |
| Use ATCL rather than ATCL+softmax (ablation) | VGG-M retrieval model · ModelNet40 | MAP: 85.35% -> 86.11% | Helped |
| Use the proposed method rather than GIFT | VGG-M retrieval model · ShapeNetCore55 perturbed, Macro | F-measure: 0.423 -> 0.404 | Hurt |
Closest papers by meaning
- Triplet-Center Loss for Multi-View 3D Object Retrieval (CVPR 2018 · similarity 0.93)
- Hard Example Generation by Texture Synthesis for Cross-domain Shape Similarity Learning (NeurIPS 2020 · similarity 0.88)
- Emphasizing 3D Properties in Recurrent Multi-View Aggregation for 3D Shape Retrieval (AAAI 2018 · similarity 0.87)
- Deep Correlated Metric Learning for Sketch-based 3D Shape Retrieval (AAAI 2017 · similarity 0.87)
- KP-RED: Exploiting Semantic Keypoints for Joint 3D Shape Retrieval and Deformation (CVPR 2024 · similarity 0.85)
- 3DCoMPaT200: Language Grounded Large-Scale 3D Vision Dataset for Compositional Recognition (NeurIPS 2024 · similarity 0.84)
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.