Comparative Analysis of Image Classification Algorithms Based on Traditional and Advanced Convolutional Neural Networks

Taimingwang Liu

2023

Abstract

This study presents a comprehensive comparative analysis of image classification algorithms across diverse datasets and distinct convolutional neural network (CNN) architectures. The datasets considered—CIFAR-10, CALTECH-101, and STL-10—embody varying complexities characteristic of real-world scenarios. They span scenarios of limited categories and low-resolution images to challenges involving diverse instances with fewer categories and high-resolution demands. The selected CNN architectures—LeNet5, VGG16, and ResNet50—exhibit varying depths and design philosophies, offering a diverse landscape for evaluation. Systematic experimentation and evaluation unveil the intricate interplay between architectural complexity and dataset characteristics. The findings underscore the pivotal role of architectural depth in addressing diverse dataset challenges. Notably, VGG16 and ResNet50 consistently outperform LeNet5 across all datasets, emphasizing the importance of deeper architectures in image classification tasks. These insights provide valuable guidance for architectural choices in image classification, ensuring alignment with specific dataset characteristics. Additionally, the study lays the foundation for future research endeavors aimed at refining architectural designs and enhancing image classification algorithm performance across various real-world scenarios.

Paper Citation

in Harvard Style

Liu T. (2023). Comparative Analysis of Image Classification Algorithms Based on Traditional and Advanced Convolutional Neural Networks. In Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-705-4, SciTePress, pages 218-223. DOI: 10.5220/0012798600003885

in Bibtex Style

@conference{daml23,
author={Taimingwang Liu},
title={Comparative Analysis of Image Classification Algorithms Based on Traditional and Advanced Convolutional Neural Networks},
booktitle={Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2023},
pages={218-223},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012798600003885},
isbn={978-989-758-705-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Comparative Analysis of Image Classification Algorithms Based on Traditional and Advanced Convolutional Neural Networks
SN - 978-989-758-705-4
AU - Liu T.
PY - 2023
SP - 218
EP - 223
DO - 10.5220/0012798600003885
PB - SciTePress