Look Closer to See Better: Recurrent Attention Convolutional Neural Network for Fine-Grained Image Recognition

By Jianlong Fu, Heliang Zheng, Tao Mei. 2017

1,285 citations recorded by the source corpus.

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

Dynamic Neural Networks: A Survey (2021, methodology): “A multi-scale CNN is built in [162], where”
This Looks Like That, Because ... Explaining Prototypes for Interpretable Image Recognition (2020, background): “This issue may also arise with other explainability methods that show or highlight part of an image, such as attention mechanisms [10, 15, 47], components [38] and other part-based models e.”
Towards Interpretable Reinforcement Learning Using Attention Augmented Agents (2019, background): “image classification and captioning [12, 13, 14, 15, 16, 17, 18, 19, 20], text classification [21, 22], generative models [6, 23, 24], object tracking [25], and reinforcement learning [26, 27, 28].”
This looks like that: deep learning for interpretable image recognition (2018, methodology): “In terms of how attention is generated, some attention models generate attention with auxiliary part-localization models trained with part annotations (e.g., [53, 52, 3, 24, 15]); other attention models generate attention with “black-box” methods – e.g., RA-CNN [7] uses another neural network (atten”
Large Scale Fine-Grained Categorization and Domain-Specific Transfer Learning (2018, methodology): “To capture subtle visual differences, visual attention [60, 16, 69] and deep metric learning [45, 10] are often used.”
Be Your Own Teacher: Improve the Performance of Convolutional Neural Networks via Self Distillation (2019, methodology): “Inspired by the intuition that neural networks should focus on critical details of input data [3], reinforcement learning and deep learning algorithms are utilized to identify the importance of pixels in the input images before they are feed into convolutional neural networks [29, 8].”
DA-GAN: Instance-Level Image Translation by Deep Attention Generative Adversarial Networks (2018, methodology): “Specifically, we jointly learn a Deep Attention Encoder (DAE) to integrate the attention mechanism [7] into the learning of the mapping function F .”
Multi-branch and Multi-scale Attention Learning for Fine-Grained Visual Categorization (2020, methodology): “Unlike RA-CNN [12], the parameters of CNN and FC in our three branches are shared.”
Counterfactual Attention Learning for Fine-Grained Visual Categorization and Re-identification (2021, neutral): “CUB Cars Aircraft RA-CNN [12] 85.3 92.5 - MA-CNN [65] 86.5 92.8 89.9 MAMC [50] 86.5 93.0 - NTS-Net [63] 87.5 93.9 91.4 WS-DAN [19] 89.4 94.5 93.0 DCL [9] 87.8 94.5 93.0 Stacked LSTM [13] Implementation Details.”
CABNet: Category Attention Block for Imbalanced Diabetic Retinopathy Grading (2020, background): “Attention mechanisms are helpful to capture the ﬁne-grained features in most computer vision tasks and have been widely used for image classiﬁcation [21], [22], object localization [23], and semantic segmentation [24], [25].”
Adaptive Triplet Model for Fine-Grained Visual Categorization (2018, methodology): “Another interesting work is RACNN [6], which uses three sub-models to train three different scale of images.”
Fine-Grained Image Analysis With Deep Learning: A Survey (2021, background): “RA-CNN iteratively generates region attention maps in a coarse to ﬁne fashion by taking previous predictions as a reference.”
High-Dimensional Sparse Cross-Modal Hashing with Fine-Grained Similarity Embedding (2021, background): “More importantly, the low-level feature similarity is less taken into consideration, which is extremely important for fine-grained retrieval [12].”
Attentive Single-Tasking of Multiple Tasks (2019, methodology): “…been used in deep learning to visualize and interpret the inner workings of CNNs [58, 68, 55], but has also been employed to improve the learned representations of convolutional networks for scale-aware semantic segmentation [5], ﬁne-grained image recognition [16] or caption generation [65, 34, 2].”
Differentiable Patch Selection for Image Recognition (2021, background): “Sequential “glimpses” : There is a long line of work that sequentially processes a sequence of patches (“glimpses”), from a network, until they settle on the most relevant ones [36, 14].”
Looking for the Devil in the Details: Learning Trilinear Attention Sampling Network for Fine-Grained Image Recognition (2019, background): “Existing attention/part-based methods [2, 7, 33, 40] try to solve this problem by learning part detectors, cropping”
Towards Robust Image Classification Using Sequential Attention Models (2019, background): “…such as question-answering [24], machine translation [6, 52], video classiﬁcation and captioning [46, 33], image classiﬁcation and captioning [37, 11, 17, 1, 55, 60, 53, 5, 57], text classiﬁcation [58, 47], generative models [42, 59, 30], object tracking [29], and reinforcement learning [10].”