Research map / Architectures and efficiency / CNN design
Relevance-CAM: Your Model Already Knows Where To Look
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 LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 2 | IoU: 0.18 -> 0.32 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 2 | Average Drop: 74.91 -> 39.02 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 2 | Average Increase: 4.45 -> 16.6 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on VGG 16 with batch normalization · ImageNet validation, layer 23 | IoU | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on VGG 16 with batch normalization · ImageNet validation, layer 23 | Average Drop: 85.43 -> 72.25 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on VGG 16 with batch normalization · ImageNet validation, layer 23 | Average Increase: 1.5 -> 3.9 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 4 | Average Drop: 23.13 -> 21.53 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on VGG 16 with batch normalization · ImageNet validation, layer 43 | Average Drop: 21.22 -> 22.42 | Hurt |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 1 | IoU: 0.12 -> 0.30 | Helped |
| Use LRP relevance scores rather than gradient weights for the CAM channels | CAM methods on ResNet 50 · ImageNet validation, layer 4 | IoU: 0.34 -> 0.34 | No clear effect |
Closest papers by meaning
- Empowering CAM-Based Methods with Capability to Generate Fine-Grained and High-Faithfulness Explanations (AAAI 2024 · similarity 0.88)
- Finer-CAM: Spotting the Difference Reveals Finer Details for Visual Explanation (CVPR 2025 · similarity 0.88)
- DD-CAM: Minimal Sufficient Explanations for Vision Models Using Delta Debugging (NeurIPS 2026 · similarity 0.86)
- CAPE: CAM as a Probabilistic Ensemble for Enhanced DNN Interpretation (CVPR 2024 · similarity 0.86)
- MoRe: Class Patch Attention Needs Regularization for Weakly Supervised Semantic Segmentation (AAAI 2025 · similarity 0.84)
- DiffCAM: Data-Driven Saliency Maps by Capturing Feature Differences (CVPR 2025 · 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.