Research map / Detection and segmentation / CNN features
Correlation Loss: Enforcing Correlation between Classification and Localization
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 |
|---|---|---|---|
| Add Correlation Loss to Sparse R-CNN using Spearman correlation | Sparse R-CNN · COCO minival | AP_C: 37.7 -> 39.3 | Helped |
| Add Correlation Loss to Sparse R-CNN using Spearman correlation | Sparse R-CNN · Cityscapes | AP: 39.0 -> 40.8 | Helped |
| Add Concordance Correlation Loss | ATSS · COCO minival | AP_C: 38.7 -> 39.8 | Helped |
| Add Concordance Correlation Loss | FoveaBox · COCO minival | AP_C: 36.4 -> 37.1 | Helped |
| Add Spearman Correlation Loss | RetinaNet · COCO minival | AP_C: 36.5 -> 37.5 | Helped |
| Add Spearman Correlation Loss | YOLACT · COCO instance segmentation | AP_C^mask: 28.3 -> 29.0 | Helped |
| Combine Correlation Loss with Rank & Sort Loss | ATSS · COCO minival | AP_C: 39.9 -> 40.2 | Helped |
| Use Correlation Loss without an auxiliary head rather than adding an auxiliary head | ATSS · COCO minival | AP_C: 39.3 vs 39.8 | Helped |
| Backpropagate Correlation Loss through only the localization head instead of only the classification head (ablation) | Sparse R-CNN · COCO minival | AP: 39.3 -> 37.5 | Hurt |
| Backpropagate Correlation Loss through both heads instead of only the classification head (ablation) | Sparse R-CNN · COCO minival | AP: 39.3 -> 38.9 | Hurt |
| Use Pearson correlation instead of Concordance correlation (ablation) | ATSS · COCO minival | AP_C: 39.8 -> 39.4 | Hurt |
| Use Pearson correlation instead of Spearman correlation (ablation) | Sparse R-CNN · COCO minival | AP_75: 42.5 -> 42.2 | Hurt |
Closest papers by meaning
- Generalized Focal Loss: Learning Qualified and Distributed Bounding Boxes for Dense Object Detection (NeurIPS 2020 · similarity 0.87)
- Searching Parameterized AP Loss for Object Detection (NeurIPS 2021 · similarity 0.87)
- RankDetNet: Delving Into Ranking Constraints for Object Detection (CVPR 2021 · similarity 0.86)
- Rethinking Classification and Localization for Object Detection (CVPR 2020 · similarity 0.86)
- Revisiting AP Loss for Dense Object Detection: Adaptive Ranking Pair Selection (CVPR 2022 · similarity 0.86)
- DR Loss: Improving Object Detection by Distributional Ranking (CVPR 2020 · 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.