Filters








4,887 Hits in 3.5 sec

Training Deep Capsule Networks with Residual Connections [article]

Josef Gugglberger, David Peer, Antonio Rodriguez-Sanchez
2021 arXiv   pre-print
In this paper, we propose a methodology to train deeper capsule networks using residual connections, which is evaluated on four datasets and three different routing algorithms.  ...  One approach to overcome such limitations would be to train deeper network architectures, as it has been done for convolutional neural networks with much increased success.  ...  We hypothesize we can be succeed at training deep capsules through the use of residual connections between capsule layers.  ... 
arXiv:2104.07393v1 fatcat:3qutr2c5incrrgglyxyevgywoy

Momentum Capsule Networks [article]

Josef Gugglberger and David Peer and Antonio Rodríguez-Sánchez
2022 arXiv   pre-print
In this paper, we provide a framework on how invertible residual building blocks can be applied to capsule networks.  ...  MoCapsNets are inspired by Momentum ResNets, a type of network that applies reversible residual building blocks.  ...  Residual learning can be applied to capsule networks as well [7] such that the training of deep capsule networks can be stabilized by using identity shortcut connections between capsule layers, which  ... 
arXiv:2201.11091v1 fatcat:q7jxwr4lm5e3bnxvohxc2hq7ea

Single Image Super Resolution via a Refined Densely Connected Inception Network

Tao Jiang, Yu Zhang, Xiaojun Wu, Gang Lu, Fei Hao, Yumei Zhang
2018 2018 25th IEEE International Conference on Image Processing (ICIP)  
Single Image Super-Resolution (SISR) has obtained unprecedented breakthrough with the development of Convolutional Neural Networks (CNN).  ...  Besides, densely connection operation is also conducted in the framework for a better use of the contextual information and feature maps.  ...  DRRN The architecture of Deep Recursive Residual Network (DRRN) [18] can be regarded as an recursive residual network, in which all the outputs of residual blocks is added to the original feature maps  ... 
doi:10.1109/icip.2018.8451441 dblp:conf/icip/JiangZWLHZ18 fatcat:6subsua5qbdlppxgqjiyrpdliu

Deep Tensor Capsule Network

Kun Sun, Liming Yuan, Haixia Xu, Xianbin Wen
2020 IEEE Access  
In order to address this issue, we propose a deep capsule network in this paper.  ...  Experimental results on CIFAR10, Fashion-MNIST, and SVHN demonstrate that the proposed deep tensor network can achieve very competitive performance compared to other state-of-the-art capsule networks.  ...  They are integrated into the residual capsule architecture, effectively improving the performance of the deep capsule network.  ... 
doi:10.1109/access.2020.2996282 fatcat:q5fvpvylcveifo7goej6tgphei

Deep Medical Image Analysis with Representation Learning and Neuromorphic Computing [article]

Neil Getty, Thomas Brettin, Dong Jin, Rick Stevens, Fangfang Xia
2020 arXiv   pre-print
This model requires less training data and outperforms both the original convolutional baseline and a previous capsule network implementation.  ...  First, we present a capsule network that explicitly learns a representation robust to rotation and affine transformation.  ...  The key component to residual networks is the introduction of a skip-connection from a previous layer's output to the next layer.  ... 
arXiv:2005.05431v1 fatcat:z3zewphpxrf4haz7w3erz3xkli

Comparisons of Deep Learning Algorithms for MNIST in Real-Time Environment

Akmaljon Palvanov, Young Im Cho
2018 International Journal of Fuzzy Logic and Intelligent Systems  
For this work, we implemented four models on the basis of unlike algorithms which are capsule network, deep residual learning model, convolutional neural network and multinomial logistic regression to  ...  Notwithstanding use of a trained model, if the model uses deep neural networks it requires much more time to make a prediction and becomes more complicated as well as memory usage also increases.  ...  ResNet As described in [24] residual network is different from a plain network with its shortcut connections.  ... 
doi:10.5391/ijfis.2018.18.2.126 fatcat:ulisydpr2bc7hegdj42bwfti4u

Fashion Image Retrieval with Capsule Networks [article]

Furkan Kınlı and BarışÖzcan and Furkan Kıraç
2019 arXiv   pre-print
In this study, we investigate in-shop clothing retrieval performance of densely-connected Capsule Networks with dynamic routing.  ...  In our design, Stacked-convolutional (SC) and Residual-connected (RC) blocks are used to form the input of capsule layers.  ...  First, we extract the features of larger-sized clothing images by more powerful methods (i.e. stacked or residual-connected convolutional layers), and forward these features to fully-connected capsules  ... 
arXiv:1908.09943v1 fatcat:iwkr4r7wezettgho3ltzrdxrky

Fashion Image Retrieval with Capsule Networks

Furkan Kinli, Baris Ozcan, Furkan Kirac
2019 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW)  
In this study, we investigate in-shop clothing retrieval performance of densely-connected Capsule Networks with dynamic routing.  ...  In our design, Stackedconvolutional (SC) and Residual-connected (RC) blocks are used to form the input of capsule layers.  ...  First, we extract the features of larger-sized clothing images by more powerful methods (i.e. stacked or residual-connected convolutional layers), and forward these features to fully-connected capsules  ... 
doi:10.1109/iccvw.2019.00376 dblp:conf/iccvw/KinliOK19 fatcat:f7xjvjt2k5dmbgmigmum5cum4a

Multiscale Feature Aggregation Capsule Neural Network for Hyperspectral Remote Sensing Image Classification

Runmin Lei, Chunju Zhang, Xueying Zhang, Jianwei Huang, Zhenxuan Li, Wencong Liu, Hao Cui
2022 Remote Sensing  
Furthermore, because deep features are generally more discriminative than shallow features, two kinds of capsule residual (CapsRES) blocks based on 3D convolutional capsule (3D-ConvCaps) layers and residual  ...  connections are proposed to increase the depth of the network and solve the limited labeled sample problem in HSI classification.  ...  Furthermore, two kinds of capsule residual blocks based on residual connections are proposed to build the deep capsule network.  ... 
doi:10.3390/rs14071652 fatcat:qwwgta6fffbsbarpcx4ycmnmx4

DE-CapsNet: A Diverse Enhanced Capsule Network with Disperse Dynamic Routing

Bohan Jia, Qiyu Huang
2020 Applied Sciences  
which uses residual convolutional layers and the position-wise dot product to build diverse enhanced primary capsules with various scales of images for complex data.  ...  Capsule Network (CapsNet) is a methodology with good prospects in visual tasks, since it can keep a stronger relationship of spatial information than Convolutional Neural Networks (CNNs).  ...  Highway Networks [10] is a deep feedforward network that provides an effective way to train networks with more than 100 layers by using bypassing paths [10] .  ... 
doi:10.3390/app10030884 fatcat:hh7l6flaivditpmguzuwus3b6e

A Novel CapsNet based Image Reconstruction and Regression Analysis

Dr. Akey Sungheetha, Dr. Rajesh Sharma R
2020 Journal of Innovative Image Processing  
Considering these issues in image classification and regression, the proposed model is designed with capsule network as an innovative method which is suitable to handle high level features.  ...  The experimental results of the proposed model are compared with conventional neural network models such as BPNN and CNN to validate the superior performance.  ...  In this a residual block is created along with a bypass layer to reduce the vanishing effects, additionally the proposed mode attains better connectivity between the layers with residual blocks to provide  ... 
doi:10.36548/jiip.2020.3.006 fatcat:3jitiemjwjgjpjz4ejo7jwg6mm

ResCaps: an improved capsule network and its application in ultrasonic image classification of thyroid papillary carcinoma

Xiongzhi Ai, Jiawei Zhuang, Yonghua Wang, Pin Wan, Yu Fu
2021 Complex & Intelligent Systems  
Capsule network (CapsNet) can effectively address tissue overlap and other problems. This paper investigates a new network model based on capsule network, which is named as ResCaps network.  ...  ResCaps network uses residual modules and enhances the abstract expression of the model.  ...  The CNN modules are consistent with the convolution parameters in the residual module. It is missing that the shortcut connection in the residual module.  ... 
doi:10.1007/s40747-021-00347-4 fatcat:z5j6jotmgrbhfccfv372klj6ie

WideCaps: A Wide Attention based Capsule Network for Image Classification [article]

S J Pawan, Rishi Sharma, Hemanth Sai Ram Reddy, M Vani, Jeny Rajan
2021 arXiv   pre-print
This paper proposes a new design strategy for capsule network architecture for efficiently dealing with complex images.  ...  The capsule network has attained unprecedented success over image classification tasks with datasets such as MNIST and affNIST by encoding the characteristic features into the capsules and building the  ...  Aggregated residual transformations for deep neural networks.  ... 
arXiv:2108.03627v2 fatcat:mkmnpws4tncb7dhxywx2s3qylu

Residual + Capsule Networks (ResCap) for Simultaneous Single-Channel Overlapped Keyword Recognition

Yan Xiong, Visar Berisha, Chaitali Chakrabarti
2019 Interspeech 2019  
Results indicate that Residual + Capsule (ResCap) network shows marked improvement in recognizing overlapped speech, especially in experiments where there is a mismatch in the number of overlapped speakers  ...  We build our network by adding capsule layers to a ResNet architecture that has shown state-of-the-art performance on a traditional keyword recognition task.  ...  Another advantage of ResCap in dealing with overlapped Conclusion This work aims to use a convolutional neural network based on a deep residual network and a capsule network to solve the single-channel  ... 
doi:10.21437/interspeech.2019-2913 dblp:conf/interspeech/XiongBC19 fatcat:27r7yspd3vdwdd2rqzcqxv63ee

ANALYSIS AND VISUALIZATION OF DEFECT DETECTION USING DEEP LEARNING MODELS

Pasumarthi Sridevi, Tirumalaraju Ashrita, Chittineni Pranavi, Vattikonda Rushyanth
2022 Zenodo  
Next, recent mainstream techniques and deep-learning methods for defects are reviewed with their shapes and sizes described.  ...  First, we classify the defects of products, like bottles, toothbrushes, leather, capsules, hazelnut, screws into categories.  ...  The deep residual network appends a residual module which is based on convolutional of neural network.  ... 
doi:10.5281/zenodo.6378956 fatcat:cofyljdjdjb25n6xr5i3i4jiti
« Previous Showing results 1 — 15 out of 4,887 results