Conference · 2026
Lightweight Convolutional Models for On-Device Crop Disease Detection (DEMO)
Daniel Mensah, Lola Adeyemi, Sarah Okafor
Abstract
DEMO publication — replace with real content. We evaluate compact convolutional architectures for classifying crop leaf diseases on low-cost smartphones, showing that pruned MobileNet variants retain 95% of full-model accuracy at a fraction of the computational cost.
Citation
Mensah, D., Adeyemi, L., & Okafor, S. (2026). Lightweight Convolutional Models for On-Device Crop Disease Detection. In Proc. DEMO Int. Conf. on Applied Machine Learning.
BibTeX
@inproceedings{mensah2026lightweight,
title={Lightweight Convolutional Models for On-Device Crop Disease Detection},
author={Mensah, Daniel and Adeyemi, Lola and Okafor, Sarah},
booktitle={Proc. DEMO Int. Conf. on Applied Machine Learning},
year={2026}
}Details
- Venue
- Proceedings of the DEMO International Conference on Applied Machine Learning
- DOI
- 10.0000/demo.2026.001
- Keywords
- crop disease; mobile inference; model compression
- Associated project
- Crop Disease Detection from Leaf Imagery (DEMO)
- Research area
- Smart Agriculture