Browsing by Author "Muhammad Huzaifa Azeem, Muhammad Ahmad Dawood and Haroon Zahid Bajwa"

Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Early rice disease identification with deep learning
    (UMT, Lahore, 2023) Muhammad Huzaifa Azeem, Muhammad Ahmad Dawood and Haroon Zahid Bajwa
    Rice is among the most widely consumed food worldwide and to avoid the loss of crop and guaranteeing food security, early diagnosis of rice disease is necessary. Yet, traditional approaches to diagnosing diseases can be difficult and heavily reliant on human judgment. In this research, we will be focusing on developing a deep learning-based automated system for detecting early rice diseases which is very easy to use by end users and gives highly accurate results. We collected a dataset of rice plant images, comprising healthy plants as well as those affected by common diseases. This dataset is made by compiling multiple datasets from Kaggle. We have a total of six classes of which one is of healthy rice and the remaining are of affected ones. We will be using pre-processing techniques like data augmentation in which we will increase the size of those classes which has fewer samples thus resulting in balanced classes which will help in achieving higher accuracy as we train the model. In this system, we will use pre-trained CNN-based models and the other methods which include Fine tuning and transfer learning. The pre-trained models that we will be using in this system will be from the following: ResNet50, ResNet101, ResNet152, ResNet50V2, ResNet101V2, ResNet152V2, DenseNet121, DenseNet169, DenseNet201, VGG16, and VGG19. These pre-trained models have the capability of giving higher accuracy as they have been trained on large datasets, which results in high performance on various tasks. They all are time-saving, cost-effective, and have wide availability. In conclusion, our research purpose is to enable farmers to recognize rice disease early by using an automated deep learning-based system and take preventative measures before it's too late and the damage is irreversible.