Improving Edge-AI Image Classification Through the Use of Better Building Blocks

Lucas Mohimont, Lilian Hollard, Luiz Steffenel

2024

Abstract

Traditional CNN architectures for classification, while successful, suffer from limitations due to diminishing spatial resolution and vanishing gradients. The emergence of modular ”building blocks” offered a new approach, allowing complex feature extraction through stacked layers. Despite the popularity of models like VGG, their high parameter count restricts their use in resource-constrained environments like Edge AI. This work investigates efficient building blocks as alternatives to VGG blocks, comparing the performance of diverse blocks from well-known models alongside our proposal block. Extensive experiments across various datasets demonstrate that our proposed block surpasses established blocks like Inception v1 in terms of accuracy while requiring significantly fewer resources regarding computational cost (GFLOPs) and memory footprint (number of parameters). This showcases its potential for real-world applications in Edge AI.

Paper Citation

in Harvard Style

Mohimont L., Hollard L. and Steffenel L. (2024). Improving Edge-AI Image Classification Through the Use of Better Building Blocks. In Proceedings of the 14th International Conference on Cloud Computing and Services Science - Volume 1: CLOSER; ISBN 978-989-758-701-6, SciTePress, pages 303-310. DOI: 10.5220/0012728000003711

in Bibtex Style

@conference{closer24,
author={Lucas Mohimont and Lilian Hollard and Luiz Steffenel},
title={Improving Edge-AI Image Classification Through the Use of Better Building Blocks},
booktitle={Proceedings of the 14th International Conference on Cloud Computing and Services Science - Volume 1: CLOSER},
year={2024},
pages={303-310},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012728000003711},
isbn={978-989-758-701-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 14th International Conference on Cloud Computing and Services Science - Volume 1: CLOSER
TI - Improving Edge-AI Image Classification Through the Use of Better Building Blocks
SN - 978-989-758-701-6
AU - Mohimont L.
AU - Hollard L.
AU - Steffenel L.
PY - 2024
SP - 303
EP - 310
DO - 10.5220/0012728000003711
PB - SciTePress