Interpretability
Class-activation map
Introduced by Oquab et al. in Is Object Localization for Free? - Weakly-Supervised Learning With Convolutional Neural NetworksClass activation maps could be used to interpret the prediction decision made by the convolutional neural network (CNN).
Image source: Learning Deep Features for Discriminative Localization
Source: Is Object Localization for Free? - Weakly-Supervised Learning With Convolutional Neural Networks
Papers
| Paper | Code | Results | Date | Stars |
|---|
Tasks
| Task | Papers | Share |
|---|---|---|
| Semantic Segmentation | 60 | 15.15% |
| Weakly-Supervised Semantic Segmentation | 49 | 12.37% |
| Object Localization | 26 | 6.57% |
| Weakly-Supervised Object Localization | 19 | 4.80% |
| Image Classification | 17 | 4.29% |
| Classification | 14 | 3.54% |
| General Classification | 11 | 2.78% |
| Pseudo Label | 8 | 2.02% |
| Weakly supervised segmentation | 8 | 2.02% |
Usage Over Time
Components
| Component | Type |
|
|---|---|---|
| 🤖 No Components Found | You can add them if they exist; e.g. Mask R-CNN uses RoIAlign |
