Abstract:Fine-grained visual categorization is challenging because the subordinate categories within an entry-level category can only be distinguished by subtle discriminations. This necessitates to localize key (most discriminative) regions and extract domain-specific features alternately, since implicit to fine-grained specialization is the existence of an entry-category visual shared among all classes. Existing methods predominantly implement fine-grained categorization independently, while neglecting that patch proposal and discrimination extraction are mutually correlated and can reinforce each other in an increasingly specialized manner. In this work, we concretize the above pipeline as an Increasing Specialized Generative Adversarial Network (IS-GAN), which recursively shapes a coarse-to-fine representation. It is a three-scale framework consisting of two highlights: a three-player expert GAN at each scale for feature extraction, and a Patch Proposal Network (PPN) between two adjacent scales for target positioning. To better anatomize pixel-to-pixel correlations at various octaves, the Gaussian pyramid and Laplacian pyramid descriptions are also integrated in each GAN. The PPN zooms the areas to shift the focus on the most representative regions by taking previous prediction of classifier as a reference, whilst a finer scale network receives an amplified attended region from previous scale. Overall, IS-GAN is driven by three focal losses from GANs and a converged object-level loss. Experiments demonstrate that IS-GAN can simultaneously (1) deliver competitive categorization performance among state-of-the-arts, i.e., validation accuracy achieves 92.23% and testing accuracy achieves 90.27%, and (2) recover fine-grained textures with high Peak Signal-to-Noise Ratios (PSNRs) (32.937) and Structural Similarities (SSIMs) (0.8607) from hand-crafted and public benchmarks.

Semi-Supervised Fine-Grained Image Categorization Using Transfer Learning with Hierarchical Multi-Scale Adversarial Networks

Increasingly Specialized Generative Adversarial Network for Fine-Grained Visual Categorization.

Fine-grained Category Discovery under Coarse-grained Supervision with Hierarchical Weighted Self-contrastive Learning

Attention GANs: Unsupervised Deep Feature Learning for Aerial Scene Classification.

Building Cross-Domain Mapping Chains from Multi-CycleGAN for Hyperspectral Image Classification

Learning a Target-Dependent Classifier for Cross-Domain Semantic Segmentation: Fine-tuning Versus Meta-Learning

Semi-supervised Cross-domain Remote Sensing Scene Classification Via Category-level Feature Alignment Network

An Adversarial Domain Adaptation Network for Cross-Domain Fine-Grained Recognition

Semantic Guided Level-Category Hybrid Prediction Network for Hierarchical Image Classification.

Adversarial Transfer Learning for Cross-Domain Visual Recognition

Semisupervised Cross-Domain Remote Sensing Scene Classification via Category-Level Feature Alignment Network

Domain Adaptive Transfer Learning on Visual Attention Aware Data Augmentation for Fine-grained Visual Categorization

Progressive Adversarial Networks for Fine-Grained Domain Adaptation.

A multi-task learning model with adversarial data augmentation for classification of fine-grained images

Adversarial Fine-Grained Adaptation Network for Cross-Scene Classification.

Webly-Supervised Fine-Grained Visual Categorization Via Deep Domain Adaptation.

Category-Level Adversaries for Semantic Domain Adaptation

A Novel Quadruple Generative Adversarial Network for Semi-Supervised Categorization of Low-Resolution Images

A Survey of Fine-Grained Image Categorization

Exploiting Images for Video Recognition with Hierarchical Generative Adversarial Networks

Classes Matter: A Fine-grained Adversarial Approach to Cross-domain Semantic Segmentation