AUTHOR=Lamba Shweta , Kukreja Vinay , Rashid Junaid , Gadekallu Thippa Reddy , Kim Jungeun , Baliyan Anupam , Gupta Deepali , Saini Shilpa 

TITLE=A novel fine-tuned deep-learning-based multi-class classifier for severity of paddy leaf diseases

JOURNAL=Frontiers in Plant Science

VOLUME=Volume 14 - 2023

YEAR=2023

URL=https://www.frontiersin.org/journals/plant-science/articles/10.3389/fpls.2023.1234067

DOI=10.3389/fpls.2023.1234067

ISSN=1664-462X

ABSTRACT=Paddy leaf diseases have a catastrophic influence on the quality and quantity of paddy grain production. Detection and recognition of the intensity of various paddy infections are critical to highquality crop production. In this paper, infections in paddy leaves are considered for the identification of illness severity. The dataset referred to is of both primary and secondary means. The four online repositories used for secondary data resources are Mendeley, GitHub, Kaggle, and UCI. The size of the dataset is 4068 images. The dataset is first pre-processed by ImageDataGenerator. Then, a Generative Adversarial Network (GAN) is performed which increased the dataset size exponentially.The severity calculation of the infected leaf is done by a number of segmentation methods. To ascertain the paddy infection, a deep learning-based hybrid approach is proposed, which combines the capabilities of a Convolutional Neural Network (CNN) and Support Vector Machine (SVM). The severity levels are determined with the assistance of a domain expert. The article in question takes four different degrees of severity into account (mild, moderate, severe, and profound). The study considered three classes for the categorization of paddy leaf diseases, i.e., bacterial blight, blast, and leaf smut.The model correctly predicts the paddy disease type and intensity with 98.43% correctness measures.The loss rate is 41.25%. The findings show that the suggested method is reliable and effective for identifying four levels of severity in bacterial blight, blast, and leaf smut infections in paddy crops. The proposed model gives better performance than the existing CNN and SVM classification models.