SugarcaneShuffleNet: A Very Fast, Lightweight Convolutional Neural Network for Diagnosis of 15 Sugarcane Leaf Diseases

2025-08-23

AI Summary

Smallholder sugarcane farmers in resource-constrained regions lack lightweight, robust, and interpretable tools for field-deployable disease diagnosis. Method: We introduce SugarcaneLD-BD—the first multi-disease dataset curated under real-world Bangladeshi field conditions—and propose SugarcaneShuffleNet, a CNN architecture optimized for low-resource environments via lightweight design, transfer learning, Bayesian hyperparameter optimization, and multi-source data augmentation. We further develop SugarcaneAI, a cross-platform mobile application integrating Grad-CAM-based visual explanations and on-device inference. Results: The model achieves 98.02% accuracy and an F1-score of 0.98, with a compact size of 9.26 MB and an inference latency of 4.14 ms per image. It outperforms state-of-the-art lightweight models (e.g., MnasNet, EdgeNeXt) in parameter count, memory footprint, and speed, significantly enhancing feasibility and interpretability of real-time field diagnosis.

Technology Category

• Machine Learning: Learning on the Edge & Model Compression
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Despite progress in AI-based plant diagnostics, sugarcane farmers in low-resource regions remain vulnerable to leaf diseases due to the lack of scalable, efficient, and interpretable tools. Many deep learning models fail to generalize under real-world conditions and require substantial computational resources, limiting their use in resource-constrained regions. In this paper, we present SugarcaneLD-BD, a curated dataset for sugarcane leaf-disease classification; SugarcaneShuffleNet, an optimized lightweight model for rapid on-device diagnosis; and SugarcaneAI, a Progressive Web Application for field deployment. SugarcaneLD-BD contains 638 curated images across five classes, including four major sugarcane diseases, collected in Bangladesh under diverse field conditions and verified by expert pathologists. To enhance diversity, we combined SugarcaneLD-BD with two additional datasets, yielding a larger and more representative corpus. Our optimized model, SugarcaneShuffleNet, offers the best trade-off between speed and accuracy for real-time, on-device diagnosis. This 9.26 MB model achieved 98.02% accuracy, an F1-score of 0.98, and an average inference time of 4.14 ms per image. For comparison, we fine-tuned five other lightweight convolutional neural networks: MnasNet, EdgeNeXt, EfficientNet-Lite, MobileNet, and SqueezeNet via transfer learning and Bayesian optimization. MnasNet and EdgeNeXt achieved comparable accuracy to SugarcaneShuffleNet, but required significantly more parameters, memory, and computation, limiting their suitability for low-resource deployment. We integrate SugarcaneShuffleNet into SugarcaneAI, delivering Grad-CAM-based explanations in the field. Together, these contributions offer a diverse benchmark, efficient models for low-resource environments, and a practical tool for sugarcane disease classification. It spans varied lighting, backgrounds and devices used on-farm

Problem

• Lack of scalable, efficient tools for sugarcane disease diagnosis
• Deep learning models require excessive computational resources
• Limited generalization of AI models in real-world conditions

Innovation

• Lightweight CNN for rapid on-device diagnosis
• Combined datasets for enhanced disease diversity
• Grad-CAM integration for field interpretability