Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

An Improved Deep Convolutional Neural Network for Image-Based Apple Plant Leaf Disease Detection and Identification

An Improved Deep Convolutional Neural Network for Image-Based Apple Plant Leaf Disease Detection... Early detection and accurate diagnosis of leaf diseases can be significantly helpful in controlling their spread and improving the yield and quality of production. However, it is challenging to design and develop novel models to analyze complex, noise-contaminated leaf images and infer, preferably in real time, with high accuracy and precision. A Deep Convolutional Neural Network (DCNN) model is proposed, trained and tested from scratch on the subset of the PlantVillage dataset comprising of typical Apple leaf-diseases images. The model uses image data augmentation and image annotation techniques to enhance performance and accuracy. The proposed model was compared with AlexNet, VGG-16, InceptionV3, MobileNetV2, ResNet50, and DenseNet121. It achieved the highest overall accuracy of 99.31% in disease detection with low training time. The low testing time of 5.1 ms per image makes the proposed model suitable for real-time disease detection. Furthermore, the proposed model achieved the maximum precision, recall, F-1 score values and was better than other models on various other performance parameters. The results were validated using a Grad-CAM visualization technique that significantly enhanced the reliability of the suggested model. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of The Institution of Engineers (India):Series A Springer Journals

An Improved Deep Convolutional Neural Network for Image-Based Apple Plant Leaf Disease Detection and Identification

Loading next page...
 
/lp/springer-journals/an-improved-deep-convolutional-neural-network-for-image-based-apple-OxiT4fkWqt
Publisher
Springer Journals
Copyright
Copyright © The Institution of Engineers (India) 2022
ISSN
2250-2149
eISSN
2250-2157
DOI
10.1007/s40030-022-00668-8
Publisher site
See Article on Publisher Site

Abstract

Early detection and accurate diagnosis of leaf diseases can be significantly helpful in controlling their spread and improving the yield and quality of production. However, it is challenging to design and develop novel models to analyze complex, noise-contaminated leaf images and infer, preferably in real time, with high accuracy and precision. A Deep Convolutional Neural Network (DCNN) model is proposed, trained and tested from scratch on the subset of the PlantVillage dataset comprising of typical Apple leaf-diseases images. The model uses image data augmentation and image annotation techniques to enhance performance and accuracy. The proposed model was compared with AlexNet, VGG-16, InceptionV3, MobileNetV2, ResNet50, and DenseNet121. It achieved the highest overall accuracy of 99.31% in disease detection with low training time. The low testing time of 5.1 ms per image makes the proposed model suitable for real-time disease detection. Furthermore, the proposed model achieved the maximum precision, recall, F-1 score values and was better than other models on various other performance parameters. The results were validated using a Grad-CAM visualization technique that significantly enhanced the reliability of the suggested model.

Journal

Journal of The Institution of Engineers (India):Series ASpringer Journals

Published: Dec 1, 2022

Keywords: Adam optimizer; AlexNet; Apple plant leaf diseases; InceptionV3; SGD optimizer; VGG-16

References