Contrastively-reinforced Attention Convolutional Neural Network for Fine-grained Image Recognition
Contrastively-reinforced Attention Convolutional Neural Network for Fine-grained Image Recognition
复制标题
DOI：
--
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Dichao Liu;Yu Wang;Jien Kato;K. Mase
中科院分区：
文献类型：
--
作者：
Dichao Liu;Yu Wang;Jien Kato;K. Mase
由于类间相似性和类内差异性，细粒度视觉分类具有固有的挑战性。然而，通过对比具有相同/不同标签的图像，人类可以本能地注意到关键线索存在于某些物体中，而其他物体则可以忽略不计。受此启发，我们提出了对比强化的注意卷积神经网络（CRA-CNN），该网络强化了深度激活的注意意识。ra - cnn主要包括两部分：分类流和注意正则化流。前者对输入图像进行分类，同时提出将输入的视觉信息划分为注意和冗余。后者通过对注意进行分类并对比各种输入的注意/冗余来评估注意/冗余建议。评价信息被反向传播，迫使分类流提高其视觉注意意识，从而有助于分类。在CUB-Birds和Stanford Cars上的实验结果表明，尽管简单，但CRA-CNN的性能明显优于基线，与最先进的研究相当。
Fine-grained visual classification is inherently challenging because of its inter-class similarity and intra-class variance. However, by contrasting the images with same/different labels, a human can instinctively notice that the key clues lie in certain objects while other objects are ignorable. Inspired by this, we propose Contrastively-reinforced Attention Convolutional Neural Network (CRA-CNN), which reinforces the attention awareness of deep activations. CRA-CNN mainly contains two parts: the classification stream and attention regularization stream. The former classifies the input image and simultaneously proposes to divide the input visual information into attention and redundancy. The latter evaluates the attention/redundancy proposal by classifying the attention and contrasting the attention/redundancy of various inputs. The evaluation information is backpropa-gated and forces the classification stream to improve its awareness of visual attention, which helps classification. Experimental results on CUB-Birds and Stanford Cars show that CRA-CNN distinctly outperforms the baselines and is comparable with state-of-art studies despite its simplicity.