Diversified Visual Attention Networks for Fine-Grained Object Classification

By Bo Zhao, Xiao Wu, Jiashi Feng, Qiang Peng, Shuicheng Yan. 2016

381 citations recorded by the source corpus.

Fine-grained object classification attracts increasing attention in multimedia applications. However, it is a quite challenging problem due to the subtle interclass difference and large intraclass variation. Recently, visual attention models have been applied to automatically localize the discriminative regions of an image for better capturing critical difference, which have demonstrated promising performance. Unfortunately, without consideration of the diversity in attention process, most of existing attention models perform poorly in classifying fine-grained objects. In this paper, we propose a diversified visual attention network (DVAN) to address the problem of fine-grained object classification, which substantially relieves the dependency on strongly supervised information for learning to localize discriminative regions com-pared with attention-less models. More importantly, DVAN explicitly pursues the diversity of attention and is able to gather discriminative information to the maximal extent. Multiple attention canvases are generated to extract convolutional features for attention. An LSTM recurrent unit is employed to learn the attentiveness and discrimination of attention canvases. The proposed DVAN has the ability to attend the object from coarse to fine granularity, and a dynamic internal representation for classification is built up by incrementally combining the information from different locations and scales of the image. Extensive experiments conducted on CUB-2011, Stanford Dogs, and Stanford Cars datasets have demonstrated that the pro-posed DVAN achieves competitive performance compared to the state-of-the-art approaches, without using any prior knowledge, user interaction, or external resource in training and testing.

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

In Computer Vision and Pattern Recognition

Citations in the corpus

Residual Attention Network for Image Classification (2017, background): “The mixed nature of attention has been studied extensively in the previous literatures [34, 16, 23, 40].”
Attention mechanisms in computer vision: A survey (2021, neutral): “[97, 46, 139, 70, 128, 36, 11, 156] , [137, 143, 63, 161, 159, 6, 177, 80] , [67, 180, 10, 23, 24, 163, 155, 151] , [44, 49, 106, 64, 168, 18, 11, 36] , [48, 16, 169, 153, 174, 127, 53, 92] , [89, 21, 5, 134, 65, 171, 2, 109] , [45, 82, 157, 91, 100, 129, 146, 131] , [138, 85, 175, 76, 166, 133, 167”
An AI-based Visual Attention Model for Vehicle Make and Model Recognition (2020, methodology): “Diversified Visual Attention Network (DVAN) [9] forces the model to vary the attention regions.”
Learning to Navigate for Fine-grained Classification (2018, methodology): “[48] propose Diversified Visual Attention Network (DVAN) to explicitly pursues the diversity of attention and better gather discriminative information.”
Spot and Learn: A Maximum-Entropy Patch Sampler for Few-Shot Image Classification (2019, background): “The work with the closest motivation to ours is [31], in which they employ a soft attention model that aims to extract all the important regions of an image by minimizing the correlation between multiple attention maps and restricts the overlap across attention maps to be lower than some threshold.”
Multi-Scale Structure-Aware Network for Human Pose Estimation (2018, methodology): “The multi-scale supervision network localizes body keypoints in a way similar to the ‘attention model’ [18] used Fig.”
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.”
DIANet: Dense-and-Implicit Attention Network (2019, methodology): “Previous works use the attention mechanism in image classification via utilizing a recurrent neural network to select and process local regions at high resolution sequentially (Mnih et al. 2014; Zhao et al. 2017).”
Multi-Attention Multi-Class Constraint for Fine-grained Image Recognition (2018, methodology): “The works in [16] and [32] introduce long short-term memory [33] and reinforcement learning to attention-based classification, respectively.”
Attention-Guided Network for Ghost-Free High Dynamic Range Imaging (2019, background): “[44] proposed a diversified visual attention network to address the problem of fine-grained object classification.”
Learning Multi-attention Convolutional Neural Network for Fine-Grained Image Recognition (2017, background): “However, the heavy human involvement makes part definition and annotation expensive and subjective, which are not optimal for all fine-grained recognition tasks [3, 36].”
A survey on deep learning-based fine-grained object classification and semantic segmentation (2017, background): “A diversiﬁed visual attention network (DVAN) [43] is proposed to pursue the diversity of attention and is able to gather discriminative information to the maximal extent.”
Learning Rich Part Hierarchies With Progressive Attention Networks for Fine-Grained Image Recognition (2020, neutral): “The attention mechanism is also integrated into fine-grained tasks to obtain discriminative parts, for example, FCANs [32] proposed a reinforcement learning framework to optimally glimpse local discriminative regions, DVAN [33] explicitly pursues the diversity of attention and is able to gather disc”
Destruction and Construction Learning for Fine-Grained Image Recognition (2019, methodology): “However, these methods [7, 42, 41, 22] usually need additional network structure (e.”