Fully Convolutional Attention Localization Networks: Efficient Attention Localization for Fine-Grained Recognition

By Xiao Liu, Tian Xia, Jiang Wang, Yuanqing Lin. 2016

102 citations recorded by the source corpus.

Fine-grained recognition is challenging mainly because the inter-class differences between fine-grained classes are usually local and subtle while intra-class differences could be large due to pose variations. In order to distinguish them from intra-class variations, it is essential to zoom in on highly discriminative local regions. In this work, we introduce a reinforcement learning-based fully convolutional attention localization network to adaptively select multiple task-driven visual attention regions. We show that zooming in on the selected attention regions significantly improves the performance of fine-grained recognition. Compared to previous reinforcement learning-based models, the proposed approach is noticeably more computationally efficient during both training and testing because of its fully-convolutional architecture, and it is capable of simultaneous focusing its glimpse on multiple visual attention regions. The experiments demonstrate that the proposed method achieves notably higher classification accuracy on three benchmark fine-grained recognition datasets: Stanford Dogs, Stanford Cars, and CUB-200-2011.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Tao Mei
2. Jiebo Luo
3. Jianlong Fu
4. Liang Lin
5. Heliang Zheng
6. Jiashi Feng
7. Shuicheng Yan
8. Yizhou Yu

In Computer Vision and Pattern Recognition

1. Luc Van Gool
2. Dacheng Tao
3. Jiashi Feng
4. Yi Yang
5. Liang Lin
6. Tao Mei
7. Xiaogang Wang
8. Ming-Hsuan Yang

Citations in the corpus

• Weakly Supervised Complementary Parts Models for Fine-Grained Image Classification From the Bottom Up (2019, methodology): “In the first group, we directly fine-tune the GoogleNet pretrained on ImageNet with the input image size set to 448 x 448, which is adopted by other algorithms [11, 30, 39] in the comparison and the classification accuracy achieved is 85.”
• Attention Convolutional Binary Neural Tree for Fine-Grained Visual Categorization (2019, methodology): “FCAN [31] ResNet-50 91.3 PA-CNN [24] VGG-19 92.6 FCAN [31] ResNet-50 × 89.1 B-CNN [29] VGG-16 × 90.6 LRBP [23] VGG-16 × 90.9 HIHCA [3] VGG-16 × 91.7 Improved B-CNN [28] VGG-16 × 92.0 BoostCNN [33] VGG-16 × 92.1 KP [8] VGG-16 × 92.4 RA-CNN [12] VGG-19 × 92.5 MA-CNN [52] VGG-19 × 92.8 MAMC [38]…”
• A survey on deep learning-based fine-grained object classification and semantic segmentation (2017, background): “FCN attention [41] is a reinforcement learning-based fully convolutional attention localization network to adaptively select multiple task-driven visual attention regions.”
• See Better Before Looking Closer: Weakly Supervised Data Augmentation Network for Fine-Grained Visual Classification (2019, unknown): “VGG-19 [23] 76.7 ResNet-101 [7] 85.8 Inception-V3 [26] 88.9 NAC [22] 68.6 PC [4] 83.8 FCAN [17] 84.2 MAMC [25] 85.2 RA-CNN [5] 87.3 WS-DAN 92.2”
• Deep neural network models for computational histopathology: A survey (2019, background): “These kinds of visual attention models have been extensively explored in computer vision applications including object detection (Liu et al., 2016), image classification (Mnih et al., 2014), image captioning (Sharma et al., 2015), and action recognition (Xu et al., 2015b) tasks.”
• Object-Part Attention Model for Fine-Grained Image Classification (2017, background): “Liu et al. [29] use the fully convolutional attention to localize multiple parts to get better classification performance.”
• Look Closer to See Better: Recurrent Attention Convolutional Neural Network for Fine-Grained Image Recognition (2017, methodology): “Comparing with the two most relevant approaches DVAN [35] and FCAN [20], the relative accuracy gains are 7.1% and 3.7%, respectively.”
• Diversified Visual Attention Networks for Fine-Grained Object Classification (2016, methodology): “Therefore, many visual attention based networks have been proposed in recent years [11], [20], [21], [24]–[27] and achieved promising results.”
• Learning Multi-attention Convolutional Neural Network for Fine-Grained Image Recognition (2017, methodology): “[18] under the same experiment setting.”
• Destruction and Construction Learning for Fine-Grained Image Recognition (2019, unknown): “8 FCAN(+BBox) [17] ResNet-50 84.”
• Fine-Grained Representation Learning and Recognition by Exploiting Hierarchical Semantic Embedding (2018, methodology): “Recently, visual attention models [5, 26, 30, 42] were intensively proposed to automatically search the informative regions and various works successfully applied this technique to FGIR [8, 17, 28, 50].”
• Context-aware attention network for image recognition (2019, methodology): “Specifically, the STCNN [41], TLAN [13], RAN [42], FCAN [14] and RACNN [30] are set as our baselines, and the experimental results and analysis will be presented in the following content.”