/Awesome-Fine-grained-Visual-Classification

Awesome Fine-grained Visual Classification

Awesome Fine-grained Visual Classification

Awesome Fine-grained Visual Classification

Survey

Papers

2021

  • [FFVT] Feature Fusion Vision Transformer. (arxiv, 2021) [paper]
  • [TPSKG] Transformer with Peak Suppression and Knowledge Guidance. (arxiv, 2021) [paper]
  • [RAMS-Trans] RAMS-Trans: Recurrent Attention Multi-scale Transformer for Fine-grained Image Recognition. (arxiv, 2021) [paper]
  • Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification. (ICCV, 2021) [paper] [code]
  • [TransFG] TransFG: A Transformer Architecture for Fine-grained Recognition. (arxiv, 2021)[paper][code]
  • Graph-Based High-Order Relation Discovery for Fine-Grained Recognition. (CVPR, 2021)[paper][code]
  • Your "Flamingo" is My "Bird": Fine-Grained, or Not (CVPR, 2021)[paper]
  • Discrimination-Aware Mechanism for Fine-Grained Representation Learning (CVPR, 2021)[paper]
  • Neural Prototype Trees for Interpretable Fine-Grained Image Recognition (CVPR, 2021) [paper]
  • Context-aware Attentional Pooling (CAP) for Fine-grained Visual Classification (AAAI, 2021) [paper]
  • Intra-class Part Swapping for Fine-Grained Image Classification (WACV, 2021) [paper]

2020

  • Interpretable and Accurate Fine-grained Recognition via Region Grouping (CVPR, 2020) [paper]
  • [LIO] Look-into-Object: Self-supervised Structure Modeling for Object Recognition (CVPR, 2020) [paper][code]
  • Weakly Supervised Fine-Grained Image Classification via Guassian Mixture Model Oriented Discriminative Learning [paper] [video]
  • Attention Convolutional Binary Neural Tree for Fine-Grained Visual Categorization (CVPR, 2020) [paper][code]
  • [CIN] Channel Interaction Networks for Fine-Grained Image Categorization (AAAI, 2020) [paper]
  • Graph-propagation based Correlation Learning for Weakly Supervised Fine-grained Image Classification (AAAI, 2020)
  • [FDL] Filtration and Distillation: Enhancing Region Attention for Fine-Grained Visual Categorization (AAAI, 2020) [paper]
  • [API-Net] Learning Attentive Pairwise Interaction for Fine-Grained Classification (AAAI, 2020) [paper]
  • [PMG] Fine-Grained Visual Classification via Progressive Multi-Granularity Training of Jigsaw Patches (ECCV, 2020)[paper]
  • [MC-loss] The Devil is in the Channels Mutual-Channel Loss for Fine-Grained Image Classification (TIP, 2020) [paper] [code]

2019

  • [TASN] Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-grained Image Recognition (CVPR, 2019) [paper]
  • Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification from the Bottom Up (CVPR, 2019)[paper]
  • [Cross-X] Cross-X Learning for Fine-Grained Visual Categorization (ICCV, 2019) [paper]
  • [DCL] Destruction and Construction Learning for Fine-grained Image Recognition (CVPR, 2019) [paper]
  • [S3N] Selective Sparse Sampling for Fine-grained Image Recognition (ICCV, 2019) [paper](https://github.com/Yao-DD/S3N "code")]
  • [MGE-CNN] Learning a Mixture of Granularity-Specific Experts for Fine-GrainedCategorization (ICCV, 2019)[paper]

2018

  • [MAMC] Multi-Attention Multi-Class Constraint forFine-grained Image Recognition (ECCV, 2018)[paper]
  • [PC] Pairwise Confusion for Fine-Grained Visual Classification (ECCV, 2018) [paper]
  • [WSBAN] Weakly Supervised Bilinear Attention Network for Fine-Grained Visual Classification (unknown, 2018) [paper]
  • [NTS-Net] Learning to Navigate for Fine-grained Classification (ECCV, 2018) [paper] [code]

2017

  • [RA-CNN] Look Closer to See Better: Recurrent Attention Convolutional Neural Networkfor Fine-grained Image Recognition (CVPR, 2017) [paper]
  • [MA-CNN] Learning Multi-Attention Convolutional Neural Network for Fine-GrainedImage Recognition (ICCV, 2017) [paper]

Paper Summary

By localize and rescale techniques

  • [RA-CNN] Look Closer to See Better: Recurrent Attention Convolutional Neural Networkfor Fine-grained Image Recognition (CVPR, 2017) [paper]
  • [MA-CNN] Learning Multi-Attention Convolutional Neural Network for Fine-GrainedImage Recognition (ICCV, 2017) [paper]
  • [NTS-Net] Learning to Navigate for Fine-grained Classification (ECCV, 2018) [paper] [code]
  • [MGE-CNN] Learning a Mixture of Granularity-Specific Experts for Fine-GrainedCategorization (ICCV, 2019)[paper]
  • [FDL] Filtration and Distillation: Enhancing Region Attention for Fine-Grained Visual Categorization (AAAI, 2020) [paper]

By metric learning

  • [MAMC] Multi-Attention Multi-Class Constraint forFine-grained Image Recognition (ECCV, 2018)[paper]
  • [PC] Pairwise Confusion for Fine-Grained Visual Classification (ECCV, 2018) [paper]
  • [CIN] Channel Interaction Networks for Fine-Grained Image Categorization (AAAI, 2020) [paper]
  • [API-Net] Learning Attentive Pairwise Interaction for Fine-Grained Classification (AAAI, 2020) [paper]

By Attention-based methods

  • [RA-CNN] Look Closer to See Better: Recurrent Attention Convolutional Neural Networkfor Fine-grained Image Recognition (CVPR, 2017) [paper]
  • [MA-CNN] Learning Multi-Attention Convolutional Neural Network for Fine-GrainedImage Recognition (ICCV, 2017) [paper]
  • [WSBAN] Weakly Supervised Bilinear Attention Network for Fine-Grained Visual Classification (unknown, 2018) [paper]
  • [Cross-X] Cross-X Learning for Fine-Grained Visual Categorization (ICCV, 2019) [paper]
  • [TASN] Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-grained Image Recognition (CVPR, 2019) [paper]
  • Attention Convolutional Binary Neural Tree for Fine-Grained Visual Categorization (CVPR, 2020) [paper][code]

Transformer-based methods

  • [TransFG] TransFG: A Transformer Architecture for Fine-grained Recognition. (arxiv, 2021)[paper][code]
  • [FFVT] Feature Fusion Vision Transformer. (arxiv, 2021) [paper]
  • [TPSKG] Transformer with Peak Suppression and Knowledge Guidance. (arxiv, 2021) [paper]
  • [RAMS-Trans] RAMS-Trans: Recurrent Attention Multi-scale Transformer for Fine-grained Image Recognition. (arxiv, 2021) [paper]

General Useful Mechanism

  • Multi-Level. (e.g., PMG / Cross-X / MGE-CNN)
  • Multi-Scale. (e.g., RA-CNN / MGE-CNN / NTS-Net/ TransFG (overlap-split) )

Recognition leaderboard

Method Backbone CUB(%) CAR(%) AIR(%) DOG(%)
RA-CNN VGG19 85.3 92.5 88.4 87.3
MA-CNN VGG19 86.5 92.8 89.9 -
MAMC ResNet101 86.5 93.0 - 85.2
PC DenseNet161 86.9 92.9 89.2 83.8
FDL DenseNet161 89.1 94.0 - 84.9
NTS-Net ResNet50 87.5 93.9 91.4 -
Cross-X ResNet50 87.7 94.6 - 88.9
S3N ResNet50 88.5 94.7 92.8 -
DCL ResNet50 87.8 94.5 93.0 -
TASN ResNet50 87.9 93.8 - -
PMG ResNet50 89.6 95.1 93.4 -
CIN ResNet50 88.1 94.5 92.8 -
API-Net DenseNet161 90.0 95.3 93.9 89.4
LIO ResNet50 88.0 94.5 92.7 -
TransFG ViT-B/16 91.7 94.8 - 92.3

Workshops

Challenges or Competitions

Datasets

Feel free to contact me if you find any interesting paper is missing.