Interpretable Attention Guided Network for Fine-grained Visual Classification

TL;DR

This paper introduces an Interpretable Attention Guided Network (IAGN) that enhances fine-grained visual classification by providing interpretable attention mechanisms and a progressive training process, achieving competitive results.

Contribution

The paper presents the first interpretable FGVC method with an attention-guided framework and a stage-wise knowledge distillation training mechanism.

Findings

Achieves competitive performance on standard FGVC datasets.

Provides interpretable attention regions for fine-grained classification.

Introduces a progressive training mechanism for multi-granularity feature fusion.

Abstract

Fine-grained visual classification (FGVC) is challenging but more critical than traditional classification tasks. It requires distinguishing different subcategories with the inherently subtle intra-class object variations. Previous works focus on enhancing the feature representation ability using multiple granularities and discriminative regions based on the attention strategy or bounding boxes. However, these methods highly rely on deep neural networks which lack interpretability. We propose an Interpretable Attention Guided Network (IAGN) for fine-grained visual classification. The contributions of our method include: i) an attention guided framework which can guide the network to extract discriminitive regions in an interpretable way; ii) a progressive training mechanism obtained to distill knowledge stage by stage to fuse features of various granularities; iii) the first interpretable FGVC method with a competitive performance on several standard FGVC benchmark datasets.