An Improved Deep Residual Convolutional Neural Network for Plant Leaf Disease Detection

In this research, we proposed a novel deep residual convolutional neural network with 197 layers (ResNet197) for the detection of various plant leaf diseases. Six blocks of layers were used to develop ResNet197. ResNet197 was trained and tested using a combined plant leaf disease image dataset. Scal...

Full description

Saved in:
Bibliographic Details
Published in:Computational intelligence and neuroscience 2022-09, Vol.2022, p.1-9
Main Authors: Pandian J., Arun, K., Kanchanadevi, Rajalakshmi, N.R., G.Arulkumaran
Format: Article
Language:English
Subjects:
Affine transformations
Artificial intelligence
Artificial neural networks
Crop diseases
Datasets
Evolution
Genetic transformation
Graphics processing units
Leaves
Medical imaging
Neural networks
Plant diseases
Plants
Saturation (color)
Transfer learning
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this research, we proposed a novel deep residual convolutional neural network with 197 layers (ResNet197) for the detection of various plant leaf diseases. Six blocks of layers were used to develop ResNet197. ResNet197 was trained and tested using a combined plant leaf disease image dataset. Scaling, cropping, flipping, padding, rotation, affine transformation, saturation, and hue transformation techniques were used to create the augmentation data of the plant leaf disease image dataset. The dataset consisted of 103 diseased and healthy image classes of 22 plants and 154,500 images of healthy and diseased plant leaves. The evolutionary search technique was used to optimise the layers and hyperparameter values of ResNet197. ResNet197 was trained on the combined plant leaf disease image dataset using a graphics processing unit (GPU) environment for 1000 epochs. It produced a 99.58 percentage average classification accuracy on the test dataset. The experimental results were superior to existing ResNet architectures and recent transfer learning techniques.
ISSN:1687-5265
1687-5273
DOI:10.1155/2022/5102290