Research map / Label-efficient and robust learning / Recognition and long-tail
Deep Differentiable Random Forests for Age Estimation
Wei Shen, Yilu Guo, Yan Wang, Kai Zhao, Bo Wang, Alan Yuille
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 |
|---|---|---|---|
| use DRF's regression forest instead of a classification forest (DNDF) (ablation) | DRF · MORPH Setup I | MAE: 3.32 -> 2.91 | Improved |
| add the forest structure instead of l2 regression loss (ablation) | DRF · MORPH Setup I | MAE: 3.21 -> 2.91 | Improved |
| use DRF leaf update functions instead of NRF's (ablation) | DRF · MORPH Setup I | MAE: 3.51 -> 2.91 | Improved |
| add DA for split nodes (ablation) | DLDLF · MORPH Setup I | MAE: 3.02 -> 2.94 | Improved |
| add DA for split nodes (ablation) | DRF (full) · MORPH Setup I | MAE: 2.85 -> 2.80 | Improved |
| add DA for split nodes (ablation) | DRF (tau=1, kmeans) · MORPH Setup I | MAE: 2.91 -> 2.85 | Improved |
| add kmeans initialization of leaf nodes (without leaf DA) (ablation) | DRF (T=0, tau=1) · MORPH Setup I | MAE: 6.91 -> 2.91 | Improved |
| add DA for leaf nodes (tau<1) without kmeans init (ablation) | DRF · MORPH Setup I | MAE: 2.85 -> 2.80 | Improved |
| add DA for leaf nodes (tau<1) without kmeans init, no split DA (ablation) | DRF (T=0, tau=1) · MORPH Setup I | MAE: 6.91 -> 2.85 | Improved |
| DRF vs DLDLF in MAE on MORPH Setup II | DRF vs DLDLF · MORPH Setup II | MAE: 2.19 -> 2.14 | Improved |
| DLDLF vs DRF in CS on MORPH Setup II | DLDLF vs DRF · MORPH Setup II | CS: 91.3% -> 93.0% | Mixed |
| train on small CACD val set: DRF vs DEX | DRF · CACD test, trained on val | MAE: 6.521 -> 5.630 | Improved |
Closest papers by meaning
- Deep Regression Forests for Age Estimation (CVPR 2018 · similarity 0.95)
- Hierarchical Attention-Based Age Estimation and Bias Analysis (TPAMI 2023 · similarity 0.87)
- Deep Cost-Sensitive and Order-Preserving Feature Learning for Cross-Population Age Estimation (CVPR 2018 · similarity 0.87)
- Adaptive Variance Based Label Distribution Learning For Facial Age Estimation (ECCV 2020 · similarity 0.87)
- Efficient Group-n Encoding and Decoding for Facial Age Estimation (TPAMI 2017 · similarity 0.87)
- Latent Factor Guided Convolutional Neural Networks for Age-Invariant Face Recognition (CVPR 2016 · similarity 0.87)
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.
