Deshpande et al. Res. Trends Int. J. Technol. Innov., April - June 2026, 1 (2) : 1-8
1Department of Agricultural Engineering, Tamil Nadu Agricultural University, Coimbatore, India; 2Department of Computer Science, Amrita Vishwa Vidyapeetham, Coimbatore, India
Article History
Accepted : 16 May 2026
Published : 29 Jun 2026
Publication Issue
Volume 1, Issue 2
April - June 2026
Page Number1–8
Early detection of foliar diseases in tomato crops can substantially reduce yield losses if diagnosis happens before symptoms spread. This paper fine-tunes a MobileNetV3 model pretrained on ImageNet to classify ten tomato leaf conditions, including early blight, late blight and leaf mould, from a field-collected dataset of 14,200 images. The model achieved 96.3 percent test accuracy while remaining small enough (5.8MB) to run on-device in a smartphone application, with inference time of 78 milliseconds on a mid-range Android device.
Keywords - crop disease detection, transfer learning, MobileNet, precision agriculture, plant pathology
Smallholder farmers often lack timely access to plant pathologists, and by the time visible symptoms are widely recognised, foliar diseases such as late blight can have already spread across a field, making on-device automated diagnosis attractive.
A dataset of 14,200 tomato leaf images spanning ten classes, nine disease categories plus healthy leaves, was assembled from field collection across three districts and augmented with rotation, brightness and occlusion transformations. A MobileNetV3-Large backbone pretrained on ImageNet was fine-tuned with a two-phase schedule, first freezing the backbone then unfreezing the final 40 layers.
The fine-tuned model achieved 96.3 percent accuracy on a held-out test set, with per-class F1-scores above 0.93 for all ten categories except leaf mould, which showed some confusion with early blight due to visually similar early-stage lesions. On-device inference on a mid-range Android phone averaged 78 milliseconds per image.
Transfer learning enables accurate, lightweight disease classifiers deployable directly on farmer smartphones without connectivity requirements. Future work will expand the dataset to cover additional regional cultivars and co-occurring nutrient deficiencies.
[1] Mohanty S. P. et al., Using deep learning for image-based plant disease detection, Frontiers in Plant Science, 2016. [2] Howard A. et al., Searching for MobileNetV3, ICCV, 2019. [3] Ferentinos K. P., Deep learning models for plant disease detection and diagnosis, Computers and Electronics in Agriculture, 2018.
© 2026 The Author(s). Published by IJEIA Editorial Office. This is an open access article under the Creative Commons Attribution 4.0 International License (CC BY 4.0).
Nikhil Deshpande, Swathi Reddy (2026). Machine Vision-Based Classification of Crop Diseases in Tomato Plants Using Transfer Learning. IJEIA, 1(2), 1-8.
Fatima Zahra Khan, Rajat Bindal
Emerging Engineering TechnologiesAyesha Siddiqui, Rohit Bhatnagar