Research map / Vision-language and multimodal / Human and embodied modeling
Dynamic MLP for Fine-Grained Image Classification by Leveraging Geographical and Temporal Information
Lingfeng Yang, Xiang Li, Renjie Song, Borui Zhao, Juntian Tao, Shihao Zhou, Jiajun Liang, Jian Yang
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 metadata-conditioned dynamic projection instead of image-only classification | Dynamic MLP with SK-Res2Net-101 · iNaturalist 2021 Mini | top-1 accuracy: 76.102 -> 84.694 | Improved |
| Use metadata-conditioned dynamic projection instead of the strongest listed multimodal baseline | Dynamic MLP with SK-Res2Net-101 · iNaturalist 2021 Full | top-1 accuracy: 91.137 -> 91.397 | Improved |
| Use dynamic MLP instead of matched-capacity concatenation fusion | iNaturalist 2021 Mini fusion comparison · iNaturalist 2021 Mini | Top-1: 83.915 -> 84.694 | Improved |
| Use dynamic MLP instead of matched-capacity addition fusion | iNaturalist 2021 Mini fusion comparison · iNaturalist 2021 Mini | Top-1: 84.068 -> 84.694 | Improved |
| Use dynamic MLP instead of matched-capacity multiplication fusion | iNaturalist 2021 Mini fusion comparison · iNaturalist 2021 Mini | Top-1: 81.649 -> 84.694 | Improved |
| Add concatenated image and multimodal features before embedding in both paths (ablation) | Dynamic MLP-C · iNaturalist 2021 Mini | top-1 accuracy: 84.341 -> 84.694 | Improved |
| Add deeper embedding layers (ablation) | Dynamic MLP · iNaturalist 2021 Mini | top-1 accuracy: 84.341 -> 84.406 | Improved |
| Concatenate inputs only in the image path rather than both paths (ablation) | Dynamic MLP-C · iNaturalist 2021 Mini | top-1 accuracy: 84.694 -> 84.573 | Hurt |
| Concatenate inputs only in the multimodal path rather than both paths (ablation) | Dynamic MLP-C · iNaturalist 2021 Mini | top-1 accuracy: 84.694 -> 84.443 | Hurt |
| Use two dynamic projection stages rather than one (ablation) | Dynamic MLP-A · iNaturalist 2021 Mini | top-1 accuracy: 84.186 -> 84.341 | Improved |
| Use two dynamic projection stages rather than three (ablation) | Dynamic MLP-A · iNaturalist 2021 Mini | top-1 accuracy: 84.052 -> 84.341 | Improved |
| Use two dynamic projection stages rather than four (ablation) | Dynamic MLP-A · iNaturalist 2021 Mini | top-1 accuracy: 84.100 -> 84.341 | Improved |
Closest papers by meaning
- Fine-Grained Recognition in the Wild: A Multi-Task Domain Adaptation Approach (ICCV 2017 · similarity 0.87)
- Webly Supervised Fine-Grained Recognition: Benchmark Datasets and an Approach (ICCV 2021 · similarity 0.87)
- Advancing Fine-Grained Classification by Structure and Subject Preserving Augmentation (NeurIPS 2024 · similarity 0.85)
- Multi-scale Activation, Refinement, and Aggregation: Exploring Diverse Cues for Fine-Grained Bird Recognition (AAAI 2025 · similarity 0.85)
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.85)
- Benchmarking Large Vision-Language Models on Fine-Grained Image Tasks: A Comprehensive Evaluation (ICLR 2026 · similarity 0.85)
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.
