Research map / Label-efficient and robust learning / Recognition and long-tail
Webly Supervised Fine-Grained Recognition: Benchmark Datasets and an Approach
Zeren Sun, Yazhou Yao, Xiu-Shen Wei, Yongshun Zhang, Fumin Shen, Jianxin Wu, Jian Zhang, Heng Tao Shen
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 Peer-learning instead of Co-teaching | B-CNN · WebFG-496 average | accuracy: 73.24 -> 76.46 | Improved |
| use Peer-learning instead of Decoupling | B-CNN · WebFG-496 average | accuracy: 73.84 -> 76.46 | Improved |
| use Peer-learning instead of Decoupling | B-CNN · Web-Aircraft | accuracy: 75.97 -> 74.38 | Hurt |
| use Peer-learning instead of plain fine-tuning | B-CNN · WebFG-496 average | accuracy: 66.10 -> 76.46 | Improved |
| use Peer-learning with ResNet-50 | ResNet-50 · WebiNat-5089 | accuracy: 48.23 -> 54.56 | Improved |
| add web images to training data | VGG-16 · FGVC-Aircraft | ACA: 84.8 -> 88.4 | Improved |
| add web images to training data | VGG-16 · CUB200-2011 | ACA: 77.7 -> 85.7 | Improved |
| add web images to training data | VGG-16 · Stanford Cars | ACA: 85.6 -> 92.4 | Improved |
| use NTS-Net backbone instead of B-CNN | Peer-learning · WebFG-496 average | accuracy: 76.46 -> 78.17 | Improved |
| use OPAM backbone instead of B-CNN | Peer-learning · WebFG-496 average | accuracy: 76.46 -> 75.32 | Hurt |
| apply label smoothing 0.1 | VGG-16 and VGG-19 · WebFG-496 | accuracy: smoothing parameter 0.1 | No clear effect |
| apply label smoothing 0.1 | ResNet-101 · WebFG-496 | accuracy: smoothing parameter 0.1 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- Co-teaching — Co-teaching: Robust training of deep neural networks with extremely noisy labels (NeurIPS 2018) · better on 4 of 4 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition (ECCV 2016 · similarity 0.93)
- Webly-Supervised Fine-Grained Visual Categorization via Deep Domain Adaptation (TPAMI 2016 · similarity 0.91)
- Webly Supervised Learning Meets Zero-Shot Learning: A Hybrid Approach for Fine-Grained Classification (CVPR 2018 · similarity 0.91)
- Fine-grained Classes and How to Find Them (ICML 2024 · similarity 0.88)
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.87)
- SELECT: A Large-Scale Benchmark of Data Curation Strategies for Image Classification (NeurIPS 2024 · 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.
