Research map / Label-efficient and robust learning / Recognition and long-tail
Adaptive Variance Based Label Distribution Learning For Facial Age Estimation
Xin Wen, Biying Li, Haiyun Guo, Zhiwei Liu, Guosheng Hu, Ming Tang, Jinqiao Wang
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 meta-learned adaptive variance instead of fixed variance (ablation) | KL-loss ResNet-18 label distribution · MORPH II Setting II | MAE: 2.66 -> 2.37 | Improved |
| add meta-learned adaptive variance instead of fixed variance (ablation) | KL-loss ResNet-18 label distribution · MORPH II Setting III | MAE: 2.79 -> 2.53 | Improved |
| add meta-learned adaptive variance instead of fixed variance (ablation) | KL-loss ResNet-18 label distribution · FG-NET | MAE: 2.64 -> 2.32 | Improved |
| add adaptive variance instead of best fixed variance 3 (ablation) | KL-loss ResNet-18 label distribution · MORPH II Setting I | MAE | Improved |
| increase fixed variance above 3 (ablation) | fixed-variance LDL baseline · MORPH II Setting I | MAE | Hurt |
| increase validation set from 1 to 3 images per class (ablation) | AVDL · MORPH II Setting I | MAE: 1.98 -> 1.94 | Improved |
| increase validation set from 2 to 3 images per class (ablation) | AVDL · MORPH II Setting I | MAE: 1.96 -> 1.94 | Improved |
| use AVDL instead of M-V Loss with same pretraining | AVDL · MORPH II Setting I | MAE: 2.16 -> 1.94 | Hurt |
| use AVDL instead of DHAA (10x more parameters) | AVDL · MORPH II Setting I | MAE: 1.91 -> 1.94 | Hurt |
| use AVDL instead of BridgeNet | AVDL · FG-NET | MAE: 2.56 -> 2.32 | Improved |
Closest papers by meaning
- Semi-Supervised Adaptive Label Distribution Learning for Facial Age Estimation (AAAI 2017 · similarity 0.91)
- Mean-Variance Loss for Deep Age Estimation From a Face (CVPR 2018 · similarity 0.88)
- Distribution Cognisant Loss for Cross-Database Facial Age Estimation With Sensitivity Analysis (TPAMI 2020 · similarity 0.88)
- Hierarchical Attention-Based Age Estimation and Bias Analysis (TPAMI 2023 · similarity 0.87)
- A Call to Reflect on Evaluation Practices for Age Estimation: Comparative Analysis of the State-of-the-Art and a Unified Benchmark (CVPR 2024 · similarity 0.85)
- From Cradle to Cane: A Two-Pass Framework for High-Fidelity Lifespan Face Aging (NeurIPS 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.
