Research map / Label-efficient and robust learning / Recognition and long-tail
Fine-Grained Visual Categorization using Meta-Learning Optimization with Sample Selection of Auxiliary Data
Yabin Zhang, Hui Tang, Kui Jia
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 |
|---|---|---|---|
| replace fine-tuning with joint training on auxiliary data | ResNet-34 · CUB-200-2011 (aux: ImageNet Subset) | accuracy: 73.5 -> 74.5 | Improved |
| replace joint training with regularized meta-learning objective (MetaFGNet) (ablation) | ResNet-34 · CUB-200-2011 (aux: ImageNet Subset) | accuracy: 74.5 -> 75.0 | Improved |
| add sample selection of auxiliary data (ablation) | MetaFGNet · CUB-200-2011 (aux: ImageNet Subset) | accuracy: 75.0 -> 75.3 | Improved |
| add sample selection of auxiliary data (ablation) | Joint training · CUB-200-2011 (aux: ImageNet Subset) | accuracy: 74.5 -> 75.0 | Improved |
| replace joint training with regularized meta-learning objective (MetaFGNet) (ablation) | ResNet-34 · CUB-200-2011 (aux: ImageNet) | accuracy: 78.8 -> 79.5 | Improved |
| add sample selection of auxiliary data (ablation) | MetaFGNet · CUB-200-2011 (aux: ImageNet) | accuracy: 79.5 -> 80.3 | Improved |
| add sample selection of auxiliary data (ablation) | Joint training · CUB-200-2011 (aux: ImageNet) | accuracy: 78.8 -> 79.4 | Improved |
| use full ImageNet instead of ImageNet Subset as auxiliary data (fine-tuning) | ResNet-34 · CUB-200-2011 | accuracy: 73.5 -> 76.8 | Improved |
| use L-Bird Subset as auxiliary data for MetaFGNet instead of ImageNet | MetaFGNet · CUB-200-2011 | accuracy: 79.5 -> 87.2 | Improved |
| add MetaFGNet meta-learning on top of fine-tuning on L-Bird Subset | ResNet-34 · CUB-200-2011 (aux: L-Bird Subset) | accuracy: 86.2 -> 87.2 | Improved |
| add sample selection of auxiliary data (ablation) | MetaFGNet · CUB-200-2011 (aux: L-Bird Subset) | accuracy: 87.2 -> 87.6 | Improved |
| add MetaFGNet with sample selection after removing dog classes from ImageNet | ResNet-34 · Stanford Dogs (ImageNet without 120 dog classes as aux) | accuracy: 69.3 -> 73.2 | Improved |
Closest papers by meaning
- An Erudite Fine-Grained Visual Classification Model (CVPR 2023 · similarity 0.90)
- Advancing Fine-Grained Classification by Structure and Subject Preserving Augmentation (NeurIPS 2024 · similarity 0.87)
- Novel Class Discovery for Ultra-Fine-Grained Visual Categorization (CVPR 2024 · similarity 0.86)
- Delving into Multimodal Prompting for Fine-Grained Visual Classification (AAAI 2024 · similarity 0.86)
- Fine-R1: Make Multi-modal LLMs Excel in Fine-Grained Visual Recognition by Chain-of-Thought Reasoning (ICLR 2026 · similarity 0.85)
- Analyzing and Boosting the Power of Fine-Grained Visual Recognition for Multi-modal Large Language Models (ICLR 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.
