Research map / Label-efficient and robust learning / Recognition and long-tail
VegFru: A Domain-Specific Dataset for Fine-Grained Visual Categorization
Saihui Hou, Yushan Feng, Zilei 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 |
|---|---|---|---|
| fuse coarse (sup-class) and fine features via a two-stream HybridNet instead of fine-label-only CBP-CNN | CBP-CNN (VGG-16) · VegFru (292 sub-classes) | top-1 mean accuracy (%): 82.21% -> 83.51% | Improved |
| fuse coarse (sup-class) and fine features via a two-stream HybridNet instead of fine-label-only CBP-CNN | CBP-CNN (VGG-16) · FGVC-Aircraft (100 sub-classes) | top-1 mean accuracy (%): 87.49% -> 88.84% | Improved |
| fuse coarse (sup-class) and fine features via a two-stream HybridNet instead of fine-label-only CBP-CNN | CBP-CNN (VGG-16) · CUB-200-2011 (200 sub-classes) | top-1 mean accuracy (%): 84.91% -> 85.78% | Improved |
| replace VGGNet with compact bilinear pooling fusion (CBP-CNN) | VGGNet · VegFru (292 sub-classes) | top-1 mean accuracy (%): 77.12% -> 82.21% | Improved |
| replace VGGNet with compact bilinear pooling fusion (CBP-CNN) | VGGNet · FGVC-Aircraft (100 sub-classes) | top-1 mean accuracy (%): 84.46% -> 87.49% | Improved |
| replace VGGNet with compact bilinear pooling fusion (CBP-CNN) | VGGNet · CUB-200-2011 (200 sub-classes) | top-1 mean accuracy (%): 72.32% -> 84.91% | Improved |
| use GoogLeNet instead of CaffeNet backbone for fine-grained sub-class classification | finetuned CaffeNet · VegFru (292 sub-classes) | top-1 mean accuracy (%): 66.40% -> 79.22% | Improved |
| use VGGNet instead of CaffeNet backbone for fine-grained sub-class classification | finetuned CaffeNet · VegFru (292 sub-classes) | top-1 mean accuracy (%): 66.40% -> 77.12% | Improved |
| use GoogLeNet instead of VGGNet on 25 sup-classes | finetuned VGGNet · VegFru (25 sup-classes) | top-1 mean accuracy (%): 82.45% -> 82.52% | No clear effect |
| use GoogLeNet instead of VGGNet on Veg200 sup-classes | finetuned VGGNet · Veg200 (15 sup-classes) | top-1 mean accuracy (%): 83.81% -> 83.50% | Hurt |
| add Coarse and Fine classifier regularization (joint finetuning) in global finetuning (ablation) | HybridNet · VegFru, Aircraft, CUB | top-1 mean accuracy (%) | Hurt |
Closest papers by meaning
- Large Scale Visual Food Recognition (TPAMI 2023 · similarity 0.87)
- Novel Class Discovery for Ultra-Fine-Grained Visual Categorization (CVPR 2024 · similarity 0.87)
- Fine-Grained Image Classification by Exploring Bipartite-Graph Labels (CVPR 2016 · similarity 0.87)
- OVFoodSeg: Elevating Open-Vocabulary Food Image Segmentation via Image-Informed Textual Representation (CVPR 2024 · similarity 0.86)
- Democratizing Fine-grained Visual Recognition with Large Language Models (ICLR 2024 · similarity 0.86)
- Exploring Domain Incremental Video Highlights Detection with the LiveFood Benchmark (AAAI 2024 · 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.
