Comparison Between CNN, ViT and CCT for Channel Frequency Response Interpretation and Application to G.Fast

被引：2

作者：

Dierickx, Philippe ^{[1
]}

Van Damme, Axel ^{[1
]}

Dupuis, Nicolas ^{[1
]}

Delaby, Olivier ^{[1
]}

机构：

[1] Nokia Bell Labs Solut Res, Dept Software & Data Syst Res SDSR Autonomous Netw, B-2018 Antwerp, Belgium

来源：

IEEE ACCESS | 2023年 / 11卷

关键词：

Transformers; Computational modeling; DSL; Frequency response; Convolutional neural networks; Data models; Channel estimation; CNN; ViT; CCT; channel frequency response;

D O I：

10.1109/ACCESS.2023.3247877

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

Convolutional Neural Networks (CNN) and more recently Visual Transformers (ViT) have been heavily used in specific areas like computer vision. Through this work, we explore and compare the CNNs and ViT models applied to a telecommunication signal, more specifically to interpret a G.fast channel frequency response. As both CNNs and ViT bring advantages, we have deepened the research by using a combination of both convolutions and transformers using Compact Convolutional Transformers (CCT) models. This study demonstrates that using transformer based models on a 1-D signal processing use case, we have significantly gained in accuracy compared to traditional convolution based models.

引用

页码：24039 / 24052

页数：14

共 13 条

[1] [Anonymous], 2019, G9701 ITUT
[2] Emerging Properties in Self-Supervised Vision Transformers
Caron, Mathilde
Touvron, Hugo
Misra, Ishan
Jegou, Herve
Mairal, Julien
Bojanowski, Piotr
Joulin, Armand
[J]. 2021 IEEE/CVF INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV 2021), 2021, : 9630 - 9640
[3] dAscoli S., 2022, J. Stat. Mech.
[4] Dosovitskiy A, 2021, Arxiv, DOI [arXiv:2010.11929, DOI 10.48550/ARXIV.2010.11929, 10.48550/arXiv.2010.11929]
[5] Dupuis N., 2019, EPO Patent, Patent No. 11146344
[6] Ericsson AB, 2012, ITU T CONTR IMPR MOD
[7] CMT: Convolutional Neural Networks Meet Vision Transformers
Guo, Jianyuan
Han, Kai
Wu, Han
Tang, Yehui
Chen, Xinghao
Wang, Yunhe
Xu, Chang
[J]. 2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR), 2022, : 12165 - 12175
[8] Hassani A, 2022, Arxiv, DOI [arXiv:2104.05704, 10.48550/ARXIV.2104.05704]
[9] Krizhevsky A, 2014, Arxiv, DOI arXiv:1404.5997
[10] Naseer M, 2021, Arxiv, DOI [arXiv:2105.10497, DOI 10.48550/ARXIV.2105.10497]

← 1 2 →