Benchmark Analysis of Representative Deep Neural Network Architectures
This paper presents an in-depth analysis of the majority of the deep neural networks (DNNs) proposed in the state of the art for image recognition. For each DNN, multiple performance indices are observed, such as recognition accuracy, model complexity, computational complexity, memory usage, and inf...
Main Authors: | , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
IEEE
2018-01-01
|
Series: | IEEE Access |
Subjects: | |
Online Access: | https://ieeexplore.ieee.org/document/8506339/ |
_version_ | 1819163039362973696 |
---|---|
author | Simone Bianco Remi Cadene Luigi Celona Paolo Napoletano |
author_facet | Simone Bianco Remi Cadene Luigi Celona Paolo Napoletano |
author_sort | Simone Bianco |
collection | DOAJ |
description | This paper presents an in-depth analysis of the majority of the deep neural networks (DNNs) proposed in the state of the art for image recognition. For each DNN, multiple performance indices are observed, such as recognition accuracy, model complexity, computational complexity, memory usage, and inference time. The behavior of such performance indices and some combinations of them are analyzed and discussed. To measure the indices, we experiment the use of DNNs on two different computer architectures, a workstation equipped with a NVIDIA Titan X Pascal, and an embedded system based on a NVIDIA Jetson TX1 board. This experimentation allows a direct comparison between DNNs running on machines with very different computational capacities. This paper is useful for researchers to have a complete view of what solutions have been explored so far and in which research directions are worth exploring in the future, and for practitioners to select the DNN architecture(s) that better fit the resource constraints of practical deployments and applications. To complete this work, all the DNNs, as well as the software used for the analysis, are available online. |
first_indexed | 2024-12-22T17:37:47Z |
format | Article |
id | doaj.art-8fc3de29ba0a481db6757a76aa14f0bb |
institution | Directory Open Access Journal |
issn | 2169-3536 |
language | English |
last_indexed | 2024-12-22T17:37:47Z |
publishDate | 2018-01-01 |
publisher | IEEE |
record_format | Article |
series | IEEE Access |
spelling | doaj.art-8fc3de29ba0a481db6757a76aa14f0bb2022-12-21T18:18:28ZengIEEEIEEE Access2169-35362018-01-016642706427710.1109/ACCESS.2018.28778908506339Benchmark Analysis of Representative Deep Neural Network ArchitecturesSimone Bianco0https://orcid.org/0000-0002-7070-1545Remi Cadene1Luigi Celona2https://orcid.org/0000-0002-5925-2646Paolo Napoletano3https://orcid.org/0000-0001-9112-0574Department of Informatics, Systems and Communication, University of Milano-Bicocca, Milan, ItalyLIP6, CNRS, Sorbonne Université, Paris, FranceDepartment of Informatics, Systems and Communication, University of Milano-Bicocca, Milan, ItalyDepartment of Informatics, Systems and Communication, University of Milano-Bicocca, Milan, ItalyThis paper presents an in-depth analysis of the majority of the deep neural networks (DNNs) proposed in the state of the art for image recognition. For each DNN, multiple performance indices are observed, such as recognition accuracy, model complexity, computational complexity, memory usage, and inference time. The behavior of such performance indices and some combinations of them are analyzed and discussed. To measure the indices, we experiment the use of DNNs on two different computer architectures, a workstation equipped with a NVIDIA Titan X Pascal, and an embedded system based on a NVIDIA Jetson TX1 board. This experimentation allows a direct comparison between DNNs running on machines with very different computational capacities. This paper is useful for researchers to have a complete view of what solutions have been explored so far and in which research directions are worth exploring in the future, and for practitioners to select the DNN architecture(s) that better fit the resource constraints of practical deployments and applications. To complete this work, all the DNNs, as well as the software used for the analysis, are available online.https://ieeexplore.ieee.org/document/8506339/Deep neural networksconvolutional neural networksimage recognition |
spellingShingle | Simone Bianco Remi Cadene Luigi Celona Paolo Napoletano Benchmark Analysis of Representative Deep Neural Network Architectures IEEE Access Deep neural networks convolutional neural networks image recognition |
title | Benchmark Analysis of Representative Deep Neural Network Architectures |
title_full | Benchmark Analysis of Representative Deep Neural Network Architectures |
title_fullStr | Benchmark Analysis of Representative Deep Neural Network Architectures |
title_full_unstemmed | Benchmark Analysis of Representative Deep Neural Network Architectures |
title_short | Benchmark Analysis of Representative Deep Neural Network Architectures |
title_sort | benchmark analysis of representative deep neural network architectures |
topic | Deep neural networks convolutional neural networks image recognition |
url | https://ieeexplore.ieee.org/document/8506339/ |
work_keys_str_mv | AT simonebianco benchmarkanalysisofrepresentativedeepneuralnetworkarchitectures AT remicadene benchmarkanalysisofrepresentativedeepneuralnetworkarchitectures AT luigicelona benchmarkanalysisofrepresentativedeepneuralnetworkarchitectures AT paolonapoletano benchmarkanalysisofrepresentativedeepneuralnetworkarchitectures |