Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Metric Learning with Spherical Embedding
Dingyi Zhang, Yingming Li, Zhongfei Zhang
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 SEC to Multi-Similarity loss | Deep metric learning · CUB200-2011 | R@1: 66.14 -> 68.79 | Improved |
| Add SEC to Multi-Similarity loss | Deep metric learning · Cars196 | R@1: 84.07 -> 85.73 | Improved |
| Add SEC to Multi-Similarity loss | Deep metric learning · SOP | R@1: 76.29 -> 78.67 | Improved |
| Add SEC to Multi-Similarity loss | Deep metric learning · In-Shop | R@1: 88.11 -> 89.87 | Improved |
| Add SEC to Cosface | Face recognition · MegaFace Challenge 1 ($10^6$ distractors) | rank-1 accuracy: 90.71 -> 91.13 | Improved |
| Add SEC to NT-Xent | SimCLR · CIFAR-10, 100 epochs | top-1 linear evaluation accuracy: 84.76 -> 86.87 | Improved |
| Add SEC to NT-Xent | SimCLR · CIFAR-100, 100 epochs | top-1 linear evaluation accuracy: 58.43 -> 61.66 | Improved |
| Use EMA to update SEC's target norm (ρ=0.01) (ablation) | Triplet Loss + SEC · Cars196 | R@1: 67.89 -> 74.57 | Improved |
| Set SEC weight η to 0.1 (ablation) | Triplet Loss + SEC · Cars196 | R@1: 60.79 -> 66.72 | Improved |
| Set SEC weight η to 0.5 (ablation) | Triplet Loss + SEC · Cars196 | R@1: 60.79 -> 67.89 | Improved |
| Increase SEC weight η to 1.5 (ablation) | Triplet Loss + SEC · Cars196 | R@1: 67.89 -> 57.72 | Hurt |
| Replace SEC with L²-reg (ablation) | Triplet Loss · Cars196 | R@1: 67.89 -> 63.02 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Cosface — CosFace: Large Margin Cosine Loss for Deep Face Recognition (CVPR 2018) · better on 2 of 2 shared results
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
- Spherical Feature Transform for Deep Metric Learning (ECCV 2020 · similarity 0.87)
- Deep Metric Learning With Angular Loss (ICCV 2017 · similarity 0.87)
- SphereFace: Deep Hypersphere Embedding for Face Recognition (CVPR 2017 · similarity 0.86)
- MetricEmbedding: Accelerate Metric Nearness by Tropical Inner Product (ICML 2025 · similarity 0.84)
- Angular Constraint Embedding via SpherePair Loss for Constrained Clustering (NeurIPS 2025 · similarity 0.83)
- Not All Classes Stand on Same Embeddings: Calibrating a Semantic Distance with Metric Tensor (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.
