An Effective Image Classification Method for Plant Diseases with Improved Channel Attention Mechanism aECAnet Based on Deep Learning

Wenqiang Yang,Ying Yuan,Donghua Zhang,Liyuan Zheng,Fuquan Nie

DOI: https://doi.org/10.3390/sym16040451

2024-04-08

Symmetry

Abstract:Since plant diseases occurring during the growth process are a significant factor leading to the decline in both yield and quality, the classification and detection of plant leaf diseases, followed by timely prevention and control measures, are crucial for safeguarding plant productivity and quality. As the traditional convolutional neural network structure cannot effectively recognize similar plant leaf diseases, in order to more accurately identify the diseases on plant leaves, this paper proposes an effective plant disease image recognition method aECA-ResNet34. This method is based on ResNet34, and in the first and the last layers of this network, respectively, we add this paper's improved aECAnet with the symmetric structure. aECA-ResNet34 is compared with different plant disease classification models on the peanut dataset constructed in this paper and the open-source PlantVillage dataset. The experimental results show that the aECA-ResNet34 model proposed in this paper has higher accuracy, better performance, and better robustness. The results show that the aECA-ResNet34 model proposed in this paper is able to recognize diseases of multiple plant leaves very accurately.

multidisciplinary sciences

What problem does this paper attempt to address?

### Problems Addressed by the Paper This paper aims to address key issues in the classification and detection of plant leaf diseases, particularly the challenges in recognizing similar plant leaf diseases. Traditional Convolutional Neural Network (CNN) structures have limitations in extracting local features, making it difficult to distinguish between similar diseased or healthy leaves. Therefore, the paper proposes an efficient image classification method based on an improved channel attention mechanism (aECAnet) called aECA-ResNet34, to enhance the accuracy of plant leaf disease recognition. #### Main Contributions: 1. **Improved Channel Attention Mechanism**: An improved channel attention mechanism, aECAnet, is proposed and its effectiveness in extracting local features from plant leaf images is validated. 2. **Novel Network Architecture**: A new network architecture combining ResNet34 with aECAnet, named aECA-ResNet34, is developed to improve model performance. 3. **Comparative Analysis**: Through a series of experiments, it is demonstrated that the proposed aECA-ResNet34 model outperforms other models, such as SENet and ECAnet, in recognizing similar plant species and similar disease symptoms. ### Experimental Results To verify the effectiveness and superiority of the aECA-ResNet34 model, researchers conducted experiments on the PlantVillage dataset and the peanut dataset. The experiments were divided into two parts: 1. **Comparison of Different Attention Mechanisms**: The aECA-ResNet34 model was compared with ResNet34 models incorporating different attention mechanisms. 2. **Comparison of Different Convolutional Neural Networks**: The aECA-ResNet34 model was compared with other convolutional neural networks. ### Datasets and Preprocessing The study used the PlantVillage dataset and the peanut dataset. The PlantVillage dataset contains 54,303 images of healthy and diseased plant leaves, divided into 39 categories, covering 14 types of plant leaves such as apple, corn, grape, bell pepper, potato, tomato, etc., and 10 different types of diseases such as leaf spot, bacterial leaf spot, early blight, late blight, etc. The specific number of images for each category is shown in Table 1. ### Conclusion By proposing a new channel attention mechanism, aECAnet, and an improved network architecture, aECA-ResNet34, this paper significantly enhances the accuracy and robustness of plant leaf disease recognition. Experimental results show that the aECA-ResNet34 model exhibits higher accuracy and better performance across multiple datasets.

An Effective Image Classification Method for Plant Diseases with Improved Channel Attention Mechanism aECAnet Based on Deep Learning

Identification of plant leaf diseases by deep learning based on channel attention and channel pruning

An effective approach for plant leaf diseases classification based on a novel DeepPlantNet deep learning model

Identification of Apple Leaf Diseases by Improved Deep Convolutional Neural Networks With an Attention Mechanism

ERCP-Net: a channel extension residual structure and adaptive channel attention mechanism for plant leaf disease classification network

An Improved Convolutional Neural Network for Plant Disease Detection Using Unmanned Aerial Vehicle Images

An Improved EfficientNetV2 Model Based on Visual Attention Mechanism: Application to Identification of Cassava Disease

Ensemble of the Deep Convolutional Network for Multiclass of Plant Disease Classification Using Leaf Images

Deep neural network-based plant protection strategy in rural garden landscape construction

Multi-Plant Disease Identification Based on Lightweight ResNet18 Model

Crop Disease Recognition Based on Modified Light-Weight CNN With Attention Mechanism

A classification and recognition model for multiple fruit tree leaf diseases

Deep Learning-Based Leaf Disease Detection in Crops Using Images for Agricultural Applications

Improved Multi-Plant Disease Recognition Method Using Deep Convolutional Neural Networks in Six Diseases of Apples and Pears

Plant leaf disease recognition based on improved SinGAN and improved ResNet34

Classification and Identification of Apple Leaf Diseases and Insect Pests Based on Improved ResNet-50 Model

A novel deep learning method for detection and classification of plant diseases

Method for Classifying Apple Leaf Diseases Based on Dual Attention and Multi-Scale Feature Extraction

Plants Disease Image Classification Based on Lightweight Convolution Neural Networks

Attention-Based Recurrent Neural Network for Plant Disease Classification

Improved AlexNet with Inception-V4 for Plant Disease Diagnosis