Attention for Fine-Grained Categorization
arXiv:1412.7054
Abstract
This paper presents experiments extending the work of Ba et al. (2014) on recurrent neural models for attention into less constrained visual environments, specifically fine-grained categorization on the Stanford Dogs data set. In this work we use an RNN of the same structure but substitute a more powerful visual network and perform large-scale pre-training of the visual network outside of the attention RNN. Most work in attention models to date focuses on tasks with toy or more constrained visual environments, whereas we present results for fine-grained categorization better than the state-of-the-art GoogLeNet classification model. We show that our model learns to direct high resolution attention to the most discriminative regions without any spatial supervision such as bounding boxes, and it is able to discriminate fine-grained dog breeds moderately well even when given only an initial low-resolution context image and narrow, inexpensive glimpses at faces and fur patterns. This and similar attention models have the major advantage of being trained end-to-end, as opposed to other current detection and recognition pipelines with hand-engineered components where information is lost. While our model is state-of-the-art, further work is needed to fully leverage the sequential input.
ICLR 2015 Workshop
References in corpus (5)
Cited by in corpus (32)
- DRAW: A Recurrent Neural Network For Image Generation
- Object-Part Attention Model for Fine-grained Image Classification
- Deep Image: Scaling up Image Recognition
- Concept Whitening for Interpretable Image Recognition
- ABC-CNN: An Attention Based Convolutional Neural Network for Visual Question Answering
- Learning to detect chest radiographs containing lung nodules using visual attention networks
- Fully Convolutional Attention Networks for Fine-Grained Recognition
- Survey on the attention based RNN model and its applications in computer vision
- Neural Activation Constellations: Unsupervised Part Model Discovery with Convolutional Networks
- Spatial Transformer Networks
- Convolutional Fine-Grained Classification with Self-Supervised Target Relation Regularization
- End-to-end Learning of Action Detection from Frame Glimpses in Videos
- Learning what to look in chest X-rays with a recurrent visual attention model
- Object Discovery From a Single Unlabeled Image by Mining Frequent Itemset With Multi-scale Features
- Learning to predict where to look in interactive environments using deep recurrent q-learning
- Dog Identification using Soft Biometrics and Neural Networks
- Agile Amulet: Real-Time Salient Object Detection with Contextual Attention
- The Unreasonable Effectiveness of Noisy Data for Fine-Grained Recognition
- Dynamic Computational Time for Visual Attention
- Localizing by Describing: Attribute-Guided Attention Localization for Fine-Grained Recognition
- Hard-Attention for Scalable Image Classification
- Embryo staging with weakly-supervised region selection and dynamically-decoded predictions
- PatternNet: Visual Pattern Mining with Deep Neural Network
- Unsupervised Part Mining for Fine-grained Image Classification
- Deep Layered Learning in MIR
- Spatial Attention Deep Net with Partial PSO for Hierarchical Hybrid Hand Pose Estimation
- Automated Weed Detection in Aerial Imagery with Context
- Pre-training Attention Mechanisms
- VERAM: View-Enhanced Recurrent Attention Model for 3D Shape Classification
- Fine-Grained Categorization via CNN-Based Automatic Extraction and Integration of Object-Level and Part-Level Features
- Adaptive and Iteratively Improving Recurrent Lateral Connections
- Recognizing Part Attributes with Insufficient Data