Improving the Accuracy of Detecting Rice Leaf Disease Using Densenet Algorithm Comparing it with Mobilenet Algorithm

Utilising machine learning methods, namely the Novel Densenet algorithm and Mobilenet, the purpose of this study effort is to evaluate the accuracy and loss in the detection of the illness that affects rice leaf surfaces. Two different methods are used on the dataset pertaining to rice leaf diseases. The purpose of this study work is to describe and construct a standard module for detecting the sickness of rice leaf. The module compares and contrasts the machine learning techniques known as Mobilenet and Novel Densenet. In order to evaluate the two groups, a total of forty sample records are collected, with twenty records belonging to each of the categories. With the assistance of clinical analysis, the sample records were evaluated; the value of alpha was found to be 0.05, the value of beta was found to be 0.5, the percentage of confidence was found to be 95, and the percentage of pretest power was found to be 80 for the website clincalc.com. On the dataset that was utilized, the machine learning algorithm Novel Densenet algorithm classifier achieves an accuracy rate of 92.91% and a loss rate of 5.52% when it comes to identifying the illness of rice leaves. On the other hand, another machine learning algorithm called Mobilenet classifier predicts the same disease at a rate of 87.52% accuracy and 15.75% loss. The independent sample t test is conducted, and the significance value (p) obtained for accuracy and loss is 0.001. This value is less than 0.05 (P < 0.05), indicating that the algorithms are statistically significant. Based on the findings of this study, it has been shown that the Novel Densenet classification algorithm is more accurate than Mobilenet when it comes to identifying diseases that affect rice leaf.