69,198 Hits in 2.6 sec

Limitation of capsule networks [article]

David Peer, Sebastian Stabinger, Antonio Rodriguez-Sanchez
2021 arXiv   pre-print
We also theoretically motivate and empirically show that this limitation affects the training of deep capsule networks negatively.  ...  Therefore, we present an incremental improvement for state-of-the-art routing algorithms that solves the aforementioned limitation and stabilizes the training of capsule networks.  ...  We conclude that the training of capsule networks is influenced by this limitation negatively and discuss the findings in the next section.  ... 
arXiv:1905.08744v4 fatcat:eszrlfx2w5b3dpngg4uqtbt4iq

Fundamental Limits of TOA/DOA and Inertial Measurement Unit-Based Wireless Capsule Endoscopy Hybrid Localization

Seongah Jeong, Joonhyuk Kang, Kaveh Pahlavan, Vahid Tarokh
2017 International Journal of Wireless Information Networks  
Here, a posterior Cramér-Rao Bound (PCRB) of the proposed TOA/DOA and IMU-based hybrid localization is derived as fundamental limits on squared position error, where the accuracies of TOA and DOA measurements  ...  In this paper, performance analysis of hybrid localization based on radio-frequency (RF) and inertial measurement unit (IMU) measurements for a single wireless capsule endoscopy (WCE) traveling the gastrointestinal  ...  to the capsule endoscopy inspection in real life.  ... 
doi:10.1007/s10776-017-0342-7 fatcat:zhycncpjnbad7alhtyejo4jdju

Comparative Performance Evaluation of VGG-16 and Capsnet using Kannada MNIST

Dr. Ramya C
2021 International Journal for Research in Applied Science and Engineering Technology  
Keywords: Capsule Networks, Deep Learning, Convolutional Neural Networks (CNNs), Kannada MNIST, VGG-16  ...  Capsule networks, which is referred to as capsNets proposed recently to overcome these shortcomings and posed to revolutionize deep learning solutions.  ...  To overcome the limitations of CNN, Sabour and Hinton et al. have recently proposed Capsule networks (Sabour et al., 2017) .  ... 
doi:10.22214/ijraset.2021.37378 fatcat:vud3onhdvbfrho33jxu7bdyspu

A Deep Learning Model with Capsules Embedded for High Resolution Image Classification

Yujuan Guo, Jingjuan Liao, Guozhuang Shen
2020 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
To solve the problem of lack of spatial information in CNNs, the Capsule network takes the form of vectors that convey location transformation information.  ...  Capsules-Unet slightly outperforms all other approaches with far fewer parameters, a reduction in parameters of over 81.8% compared with U-net and over 13.8% compared with Capsule network.  ...  In the Vaihingen dataset, the Capsules-Unet and Capsule network did not perform well when the number of samples was limited.  ... 
doi:10.1109/jstars.2020.3032672 fatcat:jvlqob2uxfhbrfdjjaky5reyke

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 proposed model achieves better retrieval efficiency of 95.4% which is much better than other neural network models.  ...  The performance of convolution neural network is much better than other neural networks, though there are few limitations in CNN model.  ... 
doi:10.36548/jiip.2020.3.006 fatcat:3jitiemjwjgjpjz4ejo7jwg6mm

ANTS: network services without the red tape

D. Wetherall, J. Guttag, D. Tennenhouse
1999 Computer  
ANTS, a new approach to deploying network services, bases interoperability on a programmable network model, not on individual networking protocols.  ...  Acknowledgments We thank the members of the Software Devices and Systems Group at MIT's Laboratory for Computer Science.  ...  This work was supported by DARPA, monitored by the Office of Naval Research, under contract N66001-96-C-8522, and by seed funding from Sun Microsystems.  ... 
doi:10.1109/2.755004 fatcat:uwwe2jmg55dhfkpaykkfab2bfy

Active network vision and reality

David Wetherall
1999 Proceedings of the seventeenth ACM symposium on Operating systems principles - SOSP '99  
of mobile code in the network.  ...  In this paper, we argue our progress towards the original vision and the difficulties that we have not yet resolved in three areas that characterize a "pure" active network: the capsule model of programmability  ...  Acknowledgments This work has benefited from the advice and assistance of many.  ... 
doi:10.1145/319151.319156 dblp:conf/sosp/Wetherall99 fatcat:wyijun2dhjcwbmdpifcwg7shli

Capsule Network Algorithm for Performance Optimization of Text Classification

Samuel Manoharan J
2021 Journal of Soft Computing Paradigm  
In regions of visual inference, optimized performance is demonstrated by capsule networks on structured data.  ...  Classification of hierarchical multi-label text is performed with a simple capsule network algorithm in this paper.  ...  However, limited applications exist when considering HMC despite the ability of capsule networks for classification into hierarchical structured categories [14] .  ... 
doi:10.36548/jscp.2021.1.001 fatcat:thebssmm3bfvrlyctvy4jheqea

Generative Adversarial Network Architectures For Image Synthesis Using Capsule Networks [article]

Yash Upadhyay, Paul Schrater
2018 arXiv   pre-print
In this paper, we propose Generative Adversarial Network (GAN) architectures that use Capsule Networks for image-synthesis.  ...  Based on the principal of positional-equivariance of features, Capsule Network's ability to encode spatial relationships between the features of the image helps it become a more powerful critic in comparison  ...  This leads to the CNN critic to learn a limited view of the manifold, thus, providing gradients to generator limited to its understanding.  ... 
arXiv:1806.03796v4 fatcat:qpm4uccua5dmxo3vfrxtj3em64

Few Shot Speaker Recognition using Deep Neural Networks [article]

Prashant Anand, Ajeet Kumar Singh, Siddharth Srivastava, Brejesh Lall
2019 arXiv   pre-print
Further, we show the effectiveness of capsule net in a few shot learning setting.  ...  To this end, we utilize an auto-encoder to learn generalized feature embeddings from class-specific embeddings obtained from capsule network.  ...  neural networks and capsule network for speaker recognition under the constraints of short and limited number of utterances. • We show that using convolutional neural network having spectrogram as input  ... 
arXiv:1904.08775v1 fatcat:bbgzqxb33ncenoe3n43tradbbq

Deep Convolutional Capsule Network for Hyperspectral Image Spectral and Spectral-Spatial Classification

Kaiqiang Zhu, Yushi Chen, Pedram Ghamisi, Xiuping Jia, Jón Atli Benediktsson
2019 Remote Sensing  
In Conv-Capsule, the number of trainable parameters is reduced compared to the original capsule, which potentially mitigates the overfitting issue when the number of available training samples is limited  ...  Furthermore, a modification of the capsule network named Conv-Capsule is proposed.  ...  Deng et al. presented a modified two-layer capsule network capable of handling a limited number of training samples for HSI classification.  ... 
doi:10.3390/rs11030223 fatcat:gvvl7hnzxjduxjepnjq3lddu6i

Introducing new Internet services: why and how

D. Wetherall, D. Legedza, J. Guttag
1998 IEEE Network  
Active networks permit applications to inject programs into the nodes of local and, more importantly, wide area networks.  ...  We explore the design of active networks by presenting a novel architecture, ants, that adds extensibility at the network layer and allows for incremental deployment of active nodes within the Internet  ...  We also wish to thank members of the wider active network community at the University of Arizona, Georgia Tech, USC ISI, Columbia University a n d U n iversity o f P ennsylvania for their support and assistance  ... 
doi:10.1109/65.690955 fatcat:goin2icjufe3xptk7w3skvzcpq

Hyperspectral Remote Sensing Image Classification Using Deep Convolutional Capsule Network

Runmin Lei, Chunju Zhang, Wencong Liu, Lei Zhang, Xueying Zhang, Yucheng Yang, Jianwei Huang, Zhenxuan Li, Zhiyi Zhou
2021 IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing  
Recently, a capsule network (CapsNet) was introduced to boost the performance of CNNs, marking a remarkable progress in the field of HSI classification.  ...  Index Terms-Capsule neural network, convolutional neural network (CNN), hyperspectral image classification.  ...  The original decoder network results in overfitting when the number of training samples is limited owing to a large number of trainable parameters.  ... 
doi:10.1109/jstars.2021.3101511 fatcat:rq7csb56cnepbojdaz2u2xcneu

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.  ...  To achieve this, we propose Triplet-based design of Capsule Network architecture with two different feature extraction methods.  ...  Therefore, within limited computational resources, these techniques are not yet applied to our models to boost the overall performance of our Capsule Network designs, and left as future research ideas.  ... 
arXiv:1908.09943v1 fatcat:iwkr4r7wezettgho3ltzrdxrky

3D-UCaps: 3D Capsules Unet for Volumetric Image Segmentation [article]

Tan Nguyen, Binh-Son Hua, Ngan Le
2022 arXiv   pre-print
Capsule network is a data-efficient network design proposed to overcome such limitations by replacing pooling layers with dynamic routing and convolutional strides, which aims to preserve the part-whole  ...  We build the concept of capsules into a CNN by designing a network with two pathways: the first pathway is encoded by 3D Capsule blocks, whereas the second pathway is decoded by 3D CNNs blocks. 3D-UCaps  ...  Disclaimer: Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.  ... 
arXiv:2203.08965v1 fatcat:62sx2axo6fbjzeraleypf2ipxy
« Previous Showing results 1 — 15 out of 69,198 results