DA-CapsNet：双注意力机制胶囊网络。

DA-CapsNet: dual attention mechanism capsule network.

机构信息

Center for Research On Leading Technology of Special Equipment, School of Mechanical and Electrical Engineering, Guangzhou University, Guangzhou, 510006, China.

School of Mechanical and Electrical Engineering, Guangzhou University, Guangzhou, 510006, China.

出版信息

Sci Rep. 2020 Jul 9;10(1):11383. doi: 10.1038/s41598-020-68453-w.

DOI:10.1038/s41598-020-68453-w

PMID:32647347

原文链接:https://pmc.ncbi.nlm.nih.gov/articles/PMC7347947/

Abstract

A capsule network (CapsNet) is a recently proposed neural network model with a new structure. The purpose of CapsNet is to form activation capsules. In this paper, our team proposes a dual attention mechanism capsule network (DA-CapsNet). In DA-CapsNet, the first layer of the attention mechanism is added after the convolution layer and is referred to as Conv-Attention; the second layer is added after the PrimaryCaps and is referred to as Caps-Attention. The experimental results show that DA-CapsNet performs better than CapsNet. For MNIST, the trained DA-CapsNet is tested in the testset, the accuracy of the DA-CapsNet is 100% after 8 epochs, compared to 25 epochs for CapsNet. For SVHN, CIFAR10, FashionMNIST, smallNORB, and COIL-20, the highest accuracy of DA-CapsNet was 3.46%, 2.52%, 1.57%, 1.33% and 1.16% higher than that of CapsNet. And the results of image reconstruction in COIL-20 show that DA-CapsNet has a more competitive performance than CapsNet.

摘要

胶囊网络（CapsNet）是一种具有新结构的新型神经网络模型。CapsNet 的目的是形成激活胶囊。在本文中，我们的团队提出了一种双重注意机制胶囊网络（DA-CapsNet）。在 DA-CapsNet 中，在卷积层之后添加了第一层注意力机制，称为 Conv-Attention；第二层添加在 PrimaryCaps 之后，称为 Caps-Attention。实验结果表明，DA-CapsNet 的性能优于 CapsNet。对于 MNIST，在测试集中对经过训练的 DA-CapsNet 进行测试，在经过 8 个周期后，DA-CapsNet 的准确率为 100%，而 CapsNet 则需要 25 个周期。对于 SVHN、CIFAR10、FashionMNIST、smallNORB 和 COIL-20，DA-CapsNet 的最高准确率比 CapsNet 高 3.46%、2.52%、1.57%、1.33%和 1.16%。并且在 COIL-20 中的图像重建结果表明，DA-CapsNet 的性能比 CapsNet 更具竞争力。