Fully Convolutional Attention Networks for Fine-Grained Recognition

Abstract

Fine-grained recognition is challenging due to its subtle local inter-class differences versus large intra-class variations such as poses. A key to address this problem is to localize discriminative parts to extract pose-invariant features. However, ground-truth part annotations can be expensive to acquire. Moreover, it is hard to define parts for many fine-grained classes. This work introduces Fully Convolutional Attention Networks (FCANs), a reinforcement learning framework to optimally glimpse local discriminative regions adaptive to different fine-grained domains. Compared to previous methods, our approach enjoys three advantages: 1) the weakly-supervised reinforcement learning procedure requires no expensive part annotations; 2) the fully-convolutional architecture speeds up both training and testing; 3) the greedy reward strategy accelerates the convergence of the learning. We demonstrate the effectiveness of our method with extensive experiments on four challenging fine-grained benchmark datasets, including CUB-200-2011, Stanford Dogs, Stanford Cars and Food-101.

References in corpus (9)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Caffe: Convolutional Architecture for Fast Feature Embedding
• Recurrent Models of Visual Attention
• Multiple Object Recognition with Visual Attention
• Bird Species Categorization Using Pose Normalized Deep Convolutional Nets
• Weakly Supervised Fine-Grained Image Categorization
• Attention for Fine-Grained Categorization
• Object-centric Sampling for Fine-grained Image Classification
• Low-rank Bilinear Pooling for Fine-Grained Classification

Cited by in corpus (29)

• Deep neural network models for computational histopathology: A survey
• This Looks Like That: Deep Learning for Interpretable Image Recognition
• Pedestrian Alignment Network for Large-scale Person Re-identification
• Object-Part Attention Model for Fine-grained Image Classification
• Bayesian Uncertainty Estimation for Batch Normalized Deep Networks
• Multi-Objective Matrix Normalization for Fine-grained Visual Recognition
• Learning Semantically Enhanced Feature for Fine-Grained Image Classification
• Part-guided Relational Transformers for Fine-grained Visual Recognition
• Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-grained Image Recognition
• Object Discovery From a Single Unlabeled Image by Mining Frequent Itemset With Multi-scale Features
• Dynamic Computational Time for Visual Attention
• Cross-X Learning for Fine-Grained Visual Categorization
• Deep Attention-guided Hashing
• Hierarchical Bilinear Pooling for Fine-Grained Visual Recognition
• Human Attention in Fine-grained Classification
• Look-into-Object: Self-supervised Structure Modeling for Object Recognition
• Knowledge-Embedded Representation Learning for Fine-Grained Image Recognition
• Deep Saliency Hashing
• Interpretable and Accurate Fine-grained Recognition via Region Grouping
• Unsupervised Part Mining for Fine-grained Image Classification
• Full-attention based Neural Architecture Search using Context Auto-regression
• Fine-Grained Representation Learning and Recognition by Exploiting Hierarchical Semantic Embedding
• Deep Imbalanced Attribute Classification using Visual Attention Aggregation
• Associating Multi-Scale Receptive Fields for Fine-grained Recognition
• Recognizing Part Attributes with Insufficient Data
• Focus Longer to See Better:Recursively Refined Attention for Fine-Grained Image Classification
• Stochastic Region Pooling: Make Attention More Expressive
• Cost-Aware Fine-Grained Recognition for IoTs Based on Sequential Fixations
• Attend and Rectify: a Gated Attention Mechanism for Fine-Grained Recovery