What type of study is this?

This is a Quantitative Study study (also classified as: Experimental Study).

October 1, 2025Open Access

A Transfer Learning-Based Efficient Model for the Detection of Plant Leaf Diseases

Key Points

The proposed model achieves an impressive testing accuracy of 97.6%, indicating high reliability in disease detection.
Using a convolutional neural network, the model effectively processes images of plant leaves to identify diseases.
Image preprocessing operations enhance data quality, vital for accurate predictions and better model performance.
The study employs transfer learning techniques on the Plant Village Dataset, improving disease recognition on unseen data.

Abstract

Plant disease control is necessary in agriculture since it can result in considerable crop yield losses. To reduce damage, quick diagnosis and categorization of plant leaf diseases is required; unfortunately, this process takes a lot of time and needs human efforts. To deal with these issues, a novel computerized approach for fast observation and categorization is required. There exist methodologies based on Deep Learning (DL) techniques that make use of an easily accessible dataset, namely The Plant Village Dataset. However, they may fail to recognize the diseases on unseen data due to less diverse feature extraction. Therefore, this research proposed a plant disease detector based on Deep Learning model using images of leaves and can identify several plant diseases. First, we perform image preprocessing operations. Second, Convolutional Neural Network (CNN) having several convolution and pooling layers is employed and the results are evaluated with existing DL models with varying hyper-parameters. After training, the model is carefully evaluated to validate the findings. We conducted several trials using the proposed model and attained testing accuracy of 97.6%

A Transfer Learning-Based Efficient Model for the Detection of Plant Leaf Diseases

Key Points

Abstract

Cite This Study