Oct 2026· International Conference on Innovative Mechanisms for Industry Applications· pp. 121-126· 0 citations· 21 references
TL;DR
A Dual-Stage Cross-Attention CNN-Transformer framework, called AgroFusionFormer, is proposed for robust plant disease detection to combine EfficientNet-B0 and Swin Transformer by a bidirectional cross-attention mechanism, where the local and global features are refined gradually.
Abstract
Plant diseases can seriously affect the productivity of agricultural production and there is a difficulty in recognising the diseases reliably in the laboratory and field due to the variation of illumination, complicated background and expanded symptom pattern. It is important to note that plant diseases can seriously affect the productivity of agricultural production and reliable disease recognition in the laboratory and field is still difficult due to illumination variation, complicated background, and spread symptom pattern. In this study, a Dual-Stage Cross-Attention CNN-Transformer framework, called AgroFusionFormer, is proposed for robust plant disease detection. It is proposed to combine EfficientNet-B0 and Swin Transformer by a bidirectional cross-attention mechanism, where the local and global features are refined gradually. Experimental evaluation on the Plant Village and PlantDoc datasets was performed with 95.84% accuracy, 95.31% precision, 94.92% recall, 95.11% F1-score and 0.978 AUC which outperforms the baselines ResNet50, EfficientNet-B0, Vision Transformer and Swin Transformer. The results of cross-environment evaluation proved the strong generalization capacity of the system and the results of the ablation analysis proved the effectiveness of the proposed fusion strategy. Moreover, the framework was found to be practical and computationally efficient with accurate localization of the diseases in images by using grad-CAM visualizations, which is useful for precision agriculture applications.
Plant diseases are a risk to agricultural productivity and require accurate and real-time detection of the disease under various field conditions. In this study, a novel YOLO-based plant disease detection framework which uses attention mechanism is proposed called YOLO-LeafNet, focusing on the multi-crop plant disease...
S. K. Mohan Raju, D. Santhakumar· International Conference on...· 0 citations
Leaf diseases are a major contributor to food and agricultural insecurity globally. The primary means of managing leaf disease through successful detection and classification of leaf diseases from field images. In this paper, we present a framework to automatically detect and classify plant leaf diseases, using a DCNN...
Anurag Gupta, T. Sardar, Sonali Bairagi et al.· International Conference on...· 0 citations
An integrated hybrid deep learning framework that combines Convolutional Neural Networks, Transformer-based self-attention, and Long Short-Term Memory networks for intelligent cotton leaf disease recognition and can serve as a scalable artificial intelligence solution for precision agriculture applications is introduce...
P. S. Gupta· Natural Resources for Human...· 0 citations
Introduction Maize crops face severe threats from various diseases and pests, significantly affecting crop quality and yield. Prompt and accurate diagnosis is essential to minimize agricultural losses. Existing models have demonstrated limitations in handling complex backgrounds, inter- and intra-class variations, and...
Timely and accurate plant disease detection is essential for sustainable agriculture and global food security. However, existing deep learning approaches still face challenges in recognizing diseases across different plant structures and imaging conditions due to variations in scale, appearance, illumination, and backg...
Betty Dewi Puspasari, I-Cheng Chang, Andy Pramono· IEEE Access· 1 citation
Agricultural pests and plant diseases are critical factors threatening global food security. Although deep learning has significantly advanced agricultural image analysis, accurate localization and recognition of multi-scale and densely distributed targets in complex natural backgrounds remain challenging. Furthermore,...
Jia-Kun He, Yue-Si Lin, Xiao-Hui Ye et al.· The International Conference...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.