Plant Disease Classifier: Detection of Dual-Crop Diseases Using Lightweight 2D CNN Architecture

Peyal, Hasibul Islam; Nahiduzzaman, Md.; Pramanik, Md. Abu Hanif; Syfullah, Md. Khalid; Shahriar, Saleh Mohammed; Sultana, Abida; Ahsan, Mominul; Haider, Julfikar; Khandakar, Amith; Chowdhury, Muhammad E. H.

doi:10.1109/access.2023.3320686

Cited by 15 publications

(9 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fig. 3 provides a comprehensive comparison across all evaluated metrics, affirming the superior performance of the proposed system over the existing methods [3,7,14]. The aggregated results highlight the effectiveness of integrating ST-GCN, DQN-AFS, and SI-MARL into a cohesive system for managing crop diseases.…”

Section: Hardware and Softwarementioning

confidence: 64%

“…Such datasets are instrumental in training and evaluating machine learning models, thereby enhancing the generalization and robustness of disease detection systems across different crop types and environmental conditions. Many of the proposed approaches are limited in scope, focusing on specific crops or diseases [3,10,11,14]. This narrow focus restricts the applicability of these methods in broader agricultural contexts and necessitates further research to generalize findings across different crop species and diseases.…”

Section: In-depth Review Of Existing Models Used For Disease Predicti...mentioning

confidence: 99%

“…DQN-AFS Configuration: The Deep Q-Network comprised three convolutional layers for feature extraction, followed by two fully connected layers for International Journal of Intelligent Engineering and Systems, Vol.17, No. 3 The model operated with a learning rate of 0.0005 and a discount factor (γ) of 0.9.SI-MARL Configuration: Each agent in the MARL framework was equipped with a policy network comprising two hidden layers. The PSO algorithm initiated with 50 particles, inertia weight (ω) ranged from 0.9 to 0.4, and personal and social learning rates (c1 and c2) were both set at 2.0.…”

Section: Model Configuration and Parametersmentioning

confidence: 99%

“…Based on this setup, the evaluation of the proposed models-Graph Attention Mechanism with Spatial-Temporal Graph Convolutional Networks (ST-GCN), Deep Q-Network with Attention-Based Feature Selection (DQN-AFS), and Swarm Intelligence-Based Multiple Agent Reinforcement Learning (SI-MARL)-was conducted using three distinct datasets: the Leaf Disease Dataset, the Temporal Crop Health Dataset, and the Treatment Efficacy Dataset Samples. These models were compared against three existing methods, referenced as [3,7,14], to assess their performance in terms of disease identification accuracy, feature selection effectiveness, and treatment recommendation precision.…”

Section: Hardware and Softwarementioning

confidence: 99%

“…International Journal of Intelligent Engineering and Systems, Vol.17, No. 3,2024 DOI: 10.22266/ijies2024.0630.55…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Design of an Iterative Method for Crop Disease Analysis Incorporating Graph Attention with Spatial-Temporal Learning and Deep Q-Networks

2024

IJIES

View full text Add to dashboard Cite

Existing approaches often fail to capture the complex spatial-temporal interactions among crop conditions, environmental factors, and disease progression, leading to suboptimal diagnosis and treatment strategies. This work introduces innovative methodologies integrating advanced machine learning techniques to overcome these limitations. Specifically, we propose a novel Graph Attention Mechanism with Spatial-Temporal Graph Convolutional Networks (ST-GCN) model to analyze crop diseases. This model uniquely combines graph attention mechanisms with spatialtemporal modeling, enabling precise identification and tracking of disease progression over time and space. This allows for high-accuracy predictions of disease presence, severity, and spatial distribution, achieving over 90% accuracy and an Intersection over Union (IoU) score of at least 0.8. Furthermore, we introduce the Deep Q-Network with Attention-Based Feature Selection (DQN-AFS) model, which innovatively applies deep Q-networks integrated with attention mechanisms to optimize feature selection in crop images. This approach significantly enhances the model's ability to discern between varying types and severity levels of crop diseases, ensuring an 85% or higher disease classification accuracy and an 80% feature selection rate in different use case scenarios. Lastly, we propose a Swarm Intelligence-Based Multiple Agent Reinforcement Learning (SI-MARL) framework for adaptive treatment recommendation. This approach demonstrates superior treatment efficacy and resource utilization efficiency compared to conventional methods.

show abstract

Section: Hardware and Softwarementioning

confidence: 64%

Section: In-depth Review Of Existing Models Used For Disease Predicti...mentioning

confidence: 99%

Section: Model Configuration and Parametersmentioning

confidence: 99%

Section: Hardware and Softwarementioning

confidence: 99%

“…International Journal of Intelligent Engineering and Systems, Vol.17, No. 3,2024 DOI: 10.22266/ijies2024.0630.55…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Design of an Iterative Method for Crop Disease Analysis Incorporating Graph Attention with Spatial-Temporal Learning and Deep Q-Networks

2024

IJIES

View full text Add to dashboard Cite

show abstract

Enhancing agriculture through real-time grape leaf disease classification via an edge device with a lightweight CNN architecture and Grad-CAM

Karim,

Goni,

Nahiduzzaman

et al. 2024

Sci Rep

View full text Add to dashboard Cite

Crop diseases can significantly affect various aspects of crop cultivation, including crop yield, quality, production costs, and crop loss. The utilization of modern technologies such as image analysis via machine learning techniques enables early and precise detection of crop diseases, hence empowering farmers to effectively manage and avoid the occurrence of crop diseases. The proposed methodology involves the use of modified MobileNetV3Large model deployed on edge device for real-time monitoring of grape leaf disease while reducing computational memory demands and ensuring satisfactory classification performance. To enhance applicability of MobileNetV3Large, custom layers consisting of two dense layers were added, each followed by a dropout layer, helped mitigate overfitting and ensured that the model remains efficient. Comparisons among other models showed that the proposed model outperformed those with an average train and test accuracy of 99.66% and 99.42%, with a precision, recall, and F1 score of approximately 99.42%. The model was deployed on an edge device (Nvidia Jetson Nano) using a custom developed GUI app and predicted from both saved and real-time data with high confidence values. Grad-CAM visualization was used to identify and represent image areas that affect the convolutional neural network (CNN) classification decision-making process with high accuracy. This research contributes to the development of plant disease classification technologies for edge devices, which have the potential to enhance the ability of autonomous farming for farmers, agronomists, and researchers to monitor and mitigate plant diseases efficiently and effectively, with a positive impact on global food security.

show abstract

FL-ToLeD: An Improved Lightweight Attention Convolutional Neural Network Model for Tomato Leaf Diseases Classification for Low-End Devices

Alnamoly,

Hady,

Abd El-Kader

et al. 2024

IEEE Access

View full text Add to dashboard Cite

The agricultural sector is still a major provider of many countries' economies, but diseases that continuously infect plants represent continuous threats to agriculture and cause massive losses to the country's economy. In this study, a lightweight convolutional neural network model called FL-ToLeD was proposed for tomato disease classification based on a soft attention mechanism with a depth-wise separable convolution layer. With a model size of 2.5 MB and 221,594 trainable parameters, FL-ToLeD achieved 99.5%, 99.10%, 99.04% for training, validation and testing accuracy respectively, and 99 % for each of precision, recall, and f1-score, it also achieved 99.90% for ROC-AUC with average inference time of 2.06924 µs. FL-ToLeD outperformed H. Ulutaş (2023) by 2.2% in terms of accuracy, precision, recall and f1-score. Additionally, it performed better than M. Agarwal (2023), Abbas (2021), and S. Verma (2020) in terms of accuracy, precision, recall, and f1-score by 8%, 2%, and 6%, respectively. It also outperformed Arshad (2023) by 4.77%, 8.92%, 35.18% and 5.11% in terms of accuracy, precision, recall and f1-score, respectively. Furthermore, FL-ToLeD is 90 times smaller than S. Verma (2020) and 2.5 times smaller than H. Ulutaş (2023) in terms of model size. All this makes FL-ToLeD more suitable for low-end devices in precision agriculture.

show abstract

Plant Disease Classifier: Detection of Dual-Crop Diseases Using Lightweight 2D CNN Architecture

Cited by 15 publications

References 31 publications

Design of an Iterative Method for Crop Disease Analysis Incorporating Graph Attention with Spatial-Temporal Learning and Deep Q-Networks

Design of an Iterative Method for Crop Disease Analysis Incorporating Graph Attention with Spatial-Temporal Learning and Deep Q-Networks

Enhancing agriculture through real-time grape leaf disease classification via an edge device with a lightweight CNN architecture and Grad-CAM

FL-ToLeD: An Improved Lightweight Attention Convolutional Neural Network Model for Tomato Leaf Diseases Classification for Low-End Devices

Contact Info

Product

Resources

About