An Enhanced Deep Learning Model for Tomato Leaf Disease Prediction

The rising need for tomato cultivation has raised concerns about detecting and diagnosing diseases affecting tomato plants. Tomato leaf disease prediction faces several challenges, including the variability of symptoms, limited availability of high-quality data, lack of standardization in disease identification and classification, and difficulty in accounting for environmental factors such as temperature and humidity. These challenges can make it challenging to develop accurate and reliable predictive models, particularly in developing countries. Detecting these diseases in the earlier stage is essential to prevent losses in yield and ensure food security. In current years, deep learning approaches have been increasingly utilized to tackle this issue. In this work, an improved deep learning model is introduced for predicting leaf diseases of tomato plants. The proposed model is based on a visual geometric group VGG-16 architecture, which has been proven effective in image classification tasks. The paper presents a Bayesian optimization method that involves fine-tuning a pre-trained VGG-16 model with a collection of images. To further increase the accuracy of the proposed work, data augmentation techniques are incorporated, such as random rotations, flips, and brightness adjustments, which increase the diversity of the training data. The diseases of tomato leaf prediction model are tested on both a publicly accessible dataset and custom dataset of tomato plant’s leaf images. The experimental results demonstrate that the VGG-16 model with Bayesian optimization approach performs well than other methods.