Journal of Shandong University(Engineering Science) ›› 2019, Vol. 49 ›› Issue (6): 1-10.doi: 10.6040/j.issn.1672-3961.0.2019.312

• The Invited Paper of the Editorial Board •     Next Articles

Review of capsule network

Jucheng YANG(),Shujie HAN,Lei MAO,Xiangzi DAI,Yarui CHEN   

  1. College of Artificial Intelligence, Tianjin University of Science and Technology, Tianjin 300457, China
  • Received:2019-06-13 Online:2019-12-20 Published:2019-12-17

RichHTML

113

PDF (PC)

1164

Abstract:

Recently capsule network with dynamic routing was the new neural network model which was considered a significant model in next generation. In recent years, much research evidenced capsule network exceptional ability to fit features. But the high computational overhead made it unable to fit complex and large datasets. Consequently, reducing computational became a research hotspot. There were two methods, including optimized capsule and optimized routing, to solve the issue. Optimized capsule was usually driven by application purpose which was designed as a model of specific classification tasks. And optimized routing was the way to improve the performance of the model from an algorithmic perspective.

Key words: capsule network, neural network, dynamic routing, optimized capsule, optimized routing

CLC Number: 

  • TP301.6

Fig.1

The structure of capsule network"

Fig.2

The structure of capsule network"

Fig.3

The orthogonal component of the feature inthe capsule subspace"

Fig.4

The structure of DCNet++"

Fig.5

The structure of PathCapsNet"

Fig.6

The structure of HitNet"

Table 1

Comparison of two capsule networks"

模型 胶囊类型 路由方法 鲁棒性 参数数量 改良方法 应用
向量胶囊网络 向量 非线性函数 弱 较少 多与卷积神经网络结合改进网络结构 多用于简单数据集
矩阵胶囊网络 矩阵 EM算法 强 较多 多改进路由算法 可用于复杂数据集

Table 2

Comparison of optimization methods"

优化方法 优化模型 优化对象 模型类型
更高特征 DCNet++[11] 胶囊 向量胶囊网络
PathCapsNet[12] 胶囊 向量胶囊网络
更精胶囊 HitNet[13] 胶囊 向量胶囊网络
CapProNet[10] 胶囊 向量胶囊网络
更少数量 SPARSECAPS[17] 胶囊 向量胶囊网络
SCNet[14] 胶囊 矩阵胶囊网络
VideoCapsuleNet[15] 胶囊 矩阵胶囊网络
S-Capsules[16] 胶囊 矩阵胶囊网络
加速路由 ICRP协议[18] 路由 向量胶囊网络
改变方法 转化为最优化问题[19] 路由 向量胶囊网络
利用KDE方法[20] 路由 矩阵胶囊网络
1 HINTON G E, KRIZHEVSKY A, WANG S D. Transforming auto-encoders[C]//International Conference on Artificial Neural Networks. Berlin, Germany: Springer, 2011: 44-51.
2 SABOUR S, FROSST N, HINTON G E. Dynamic routing between capsules[C]//Neural Information Processing Systems. California, USA: NIPS Proceeding, 2017: 3856-3866.
3 HINTON G E, SABOUR S, FROSST N. Matrix capsules with EM routing[EB/OL].(2018-04)[2019-06-13]. https://openreview.net/pdf?id=HJWLfGWRb.
4 ZHANG L, EDRAKI M, QI G J. CapProNet: deep feature learning via orthogonal projections onto capsule subspaces[C]//Neural Information Processing Systems. Montréal, Canada: NIPS Proceeding, 2018: 5814-5823.
5 PHAYE S S R, SIKKA A, DHALL A, et al. Dense and diverse capsule networks: making the capsules learn better[EB/OL].(2018-05)[2019-06-13]. https://arxiv.org/pdf/1805.04001.pdf.
6 SHAHROUDNEJAD A, AFSHAR P, PLATANIOTIS K N, et al. Improved explainability of capsule networks: Relevance path by agreement[C]//2018 IEEE Global Conference on Signal and Information Processing. California, USA: IEEE, 2018: 549-553.
7 DELIÈKGE A, CIOPPA A, VAN DROOGENBROECK M. HitNet: a neural network with capsules embedded in a Hit-or-Miss layer, extended with hybrid data augmentation and ghost capsules[EB/OL].(2018-06)[2019-06-13]. https://arxiv.org/pdf/1806.06519.pdf.
8 NEILL J O. Siamese capsule networks[EB/OL].(2018-05)[2019-06-13]. https://arxiv.org/pdf/1805.07242.pdf.
9 DUARTE K, RAWAT Y, SHAH M. Videocapsulenet: a simplified network for action detection[C]//Neural Information Processing Systems. Montréal, Canada: NIPS Proceeding, 2018: 7610-7619.
10 BAHADORI M T. Spectral capsule networks[EB/OL].(2018-02)[2019-06-13]. https://openreview.net/pdf?id=HJuMvYPaM.
11 RAWLINSON D, AHMED A, KOWADLO G. Sparse unsupervised capsules generalize better[EB/OL].(2018-04)[2019-06-13]. https://arxiv.org/pdf/1804.06094.pdf.
12 SAHU S K, KUMAR P, SINGH A P. Dynamic routing using inter capsule routing protocol between capsules[C]//2018 UKSim-AMSS 20th International Conference on Computer Modelling and Simulation. Cambridge, UK: IEEE, 2018: 1-5.
13 WANG D, LIU Q. An optimization view on dynamic routing between capsules[EB/OL].(2018-02)[2019-06-13]. https://openreview.net/pdf?id=HJjtFYJDf.
14 ZHANG S, ZHOU Q, WU X. Fast dynamic routing based on weighted kernel density estimation[C]//International Symposium on Artificial Intelligence and Robotics. Nanjing, China: Springer, 2018: 301-309.
15 MOBINY A, VAN NGUYEN H.Fast capsnet for lung cancer screening[C]//International Conference on Medical Image Computing and Computer-Assisted Intervention. Granada, Spain: Springer, 2018: 741-749.
16 KIM Y, WANG P, ZHU Y, et al. A capsule network for traffic speed prediction in complex road networks[C]//2018 Sensor Data Fusion: Trends, Solutions, Applications. Nordrhein-Westfalen, Germany: IEEE, 2018: 1-6.
17 JIMÉINEZ-SÁBNCHEZ A, ALBARQOUNI S, MATEUS D. Capsule networks against medical imaging data challenges[EB/OL]. (2018-07)[2019-06-13]. https://arxiv.org/pdf/1807.07559.pdf.
18 KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]// Neural Information Processing Systems. Lake Tahoe, USA: NIPS Proceeding, 2012: 1097-1105.
19 NGUYEN D Q, VU T, NGUYEN T D, et al. A capsule network-based embedding model for knowledge graph completion and search personalization [EB/OL].(2018-04)[2019-06-13]. https://arxiv.org/pdf/1808.04122.pdf.
20 AFSHAR P, MOHAMMADI A, PLATANIOTIS K N. Brain tumor type classification via capsule networks[C]//2018 25th IEEE International Conference on Image Processing. Athens, Greece: IEEE, 2018: 3129-3133.
21 DE JESUS D R, CUEVAS J, RIVERA W, et al. Capsule networks for protein structure classification and prediction[EB/OL].(2018-08)[2019-06-13]. https://arxiv.org/pdf/1808.07475.pdf.
22 DENG F , PU S , CHEN X , et al. Hyperspectral image classification with capsule network using limited training samples[J]. Sensors, 2018, 18 (9): 3153.
doi: 10.3390/s18093153
23 UPADHYAY Y, SCHRATER P. Generative adversarial network architectures for image synthesis using capsule networks[EB/OL].(2018-06)[2019-06-13]. https://arxiv.org/pdf/1806.03796.pdf.
24 JAISWAL A, ABDALMAGEED W, WU Y, et al. Capsulegan: generative adversarial capsule network[C]//European Conference on Computer Vision. Munich, Germany: Springer, 2018: 526-535.
25 ZHAO W, YE J, YANG M, et al. Investigating capsule networks with dynamic routing for text classification[EB/OL].(2018-05)[2019-06-13]. https://arxiv.org/pdf/1804.00538.pdf.
[1] Diankun ZHENG,Tongle XU,Zhaojie YIN,Qingmin MENG. Prediction method of tailing dam groundwater levels based on improved PSO-BP neural network [J]. Journal of Shandong University(Engineering Science), 2019, 49(3): 108-113.
[2] Yijiang HE,Junping DU,Feifei KOU,Meiyu LIANG,Wei WANG,Ang LUO. Images auto-encoding algorithm based on deep convolution neural network [J]. Journal of Shandong University(Engineering Science), 2019, 49(2): 61-66.
[3] Xiaoxiong HOU,Xinzheng XU,Jiong ZHU,Yanyan GUO. Computer aided diagnosis method for breast cancer based on AlexNet and ensemble classifiers [J]. Journal of Shandong University(Engineering Science), 2019, 49(2): 74-79.
[4] Fang GUO,Lei CHEN,Ziwen YANG. Real-time traffic prediction based on MGU for large-scale IP backbone networks [J]. Journal of Shandong University(Engineering Science), 2019, 49(2): 88-95.
[5] Ya'nan YANG,Bin XIA,Nan XIE,Wenhao YUAN. Hybrid localization algorithm based on BP neural network and multivariable Taylor series [J]. Journal of Shandong University(Engineering Science), 2019, 49(1): 36-40.
[6] Wenwen QUAN,Mingxing LIN. Algorithm of underwater target recognition based on CNN features with BOF [J]. Journal of Shandong University(Engineering Science), 2019, 49(1): 107-113.
[7] Dongdong SHEN,Fengyu ZHOU,Mengyuan LI,Shuqian WANG,Renhe GUO. Indoor wireless positioning based on ensemble deep neural network [J]. Journal of Shandong University(Engineering Science), 2018, 48(5): 95-102.
[8] Pu ZHANG,Chang LIU,Yong WANG. Suggestion sentence classification model based on feature fusion and ensemble learning [J]. Journal of Shandong University(Engineering Science), 2018, 48(5): 47-54.
[9] Mengmeng LIANG,Tao ZHOU,Yong XIA,Feifei ZHANG,Jian YANG. Lung tumor images recognition based on PSO-ConvK convolutional neural network [J]. Journal of Shandong University(Engineering Science), 2018, 48(5): 77-84.
[10] ZHANG Xianhong, ZHANG Chunrui. Image enhancement algorithm based on six dimensional feedforward neural network model [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(4): 10-19.
[11] LI Yuxin, PU Yuanyuan, XU Dan, QIAN Wenhua, LIU Hejuan. Image aesthetic quality evaluation based on embedded fine-tune deep CNN [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(3): 60-66.
[12] ZHAO Yanxia, WANG Xizhao. Multipurpose zero watermarking algorithm for color image based on SVD and DCNN [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(3): 25-33.
[13] CAO Ya, DENG Zhaohong, WANG Shitong. An radial basis function neural network model based on monotonic constraints [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(3): 127-133.
[14] XIE Zhifeng, WU Jiaping, MA Lizhuang. Chinese financial news classification method based on convolutional neural network [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(3): 34-39.
[15] HE Zhengyi, ZENG Xianhua, GUO Jiang. An ensemble method with convolutional neural network and deep belief network for gait recognition and simulation [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2018, 48(3): 88-95.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
[1] Yue Khing Toh1， XIAO Wendong2， XIE Lihua1. Wireless sensor network for distributed target tracking: practices via real test bed development[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2009, 39(1): 50 -56 .
[2] XU Li-li,JI Zhong,XIA Ji-mei . The optimum algorithm for the container loading problem with homogeneous cargoes[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2008, 38(3): 14 -17 .
[3] PAN Duo-tao,LIU Gui-ping,LIU Chang-feng . Screening of microbe producing flocculant and optimizationon its cultural conditions[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2008, 38(3): 99 -103 .
[4] ZHANG Gong-xiao,YANG Rong-hua . Synthesis and characterization of salicylaldehyde methylthiosemicarbazone Schiff base complexes[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2008, 38(3): 108 -111 .
[5] XU Xiaodan, DUAN Zhengjie, CHEN Zhongyu. The sentiment mining method based on extended sentiment dictionary and integrated features[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2014, 44(6): 15 -18 .
[6] SUN Xiang-yong . A note on total colorings of planar graphs without 4cycle and adjacent 3cycle[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2007, 37(3): 118 -121 .
[7] LI Li-ping,LI Shu-cai,XU Bang-shu,DING Wan-tao,WEI Li-yuan . Numerical optimization study on the construction of subsea tunnel
[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2008, 38(4): 63 -68 .
[8] GAO Yu, DU Ai-ling, YANG Yan-ling. Influence of zinc oxide in a phosphating bath on phosphate coating on AZ61 magnesium alloy[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2009, 39(3): 77 -80 .
[9] GAO Ming, SHI Yue-Thao, WANG Ni-Ni, SUN Feng-Zhong, PING Ya-Ming. Circumferential inflow air distributing rules in a natural draft  wet-cooling tower under crosswind conditions[J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2009, 39(3): 154 -158 .
[10] LI Chun-xiao， YUE Qin-yan， LU Lei， GAO Bao-yu， YANG Zhong-lian， SI Xiao-hui， NI Shou-qing， WANG Yuan-fang. Synthesis and application of  hydrophobically  associating  cationic  polyacrylamide    [J]. JOURNAL OF SHANDONG UNIVERSITY (ENGINEERING SCIENCE), 2008, 38(6): 99 -104 .
Copyright 2018 © Editorial office of Journal of Shandong University (Engineering Science)
Supported by Beijing Magtech Co.Ltd