Accuracy Improvement in Disease Identification of Mango Leaf Using CNN Algorithm Compared with Fuzzy Algorithm
Sivaram Chowdary M., R Puviarasi
Abstract
Aim: The main aim of this work is to measure the accuracy in the identification of mango leaf diseases using Convolutional Neural Networks (CNN) compared with Fuzzy Logic. Materials and methods: The data set contains 10 images collected from the seed buzz website and these images are used for training and testing the predictive model in MATLAB. Statistical analysis is done using SPSS software. In the SPSS tool, the measured accuracy of CNN is compared with the Fuzzy model accuracy. Result: The proposed system using CNN achieved high accuracy of 95.2%, whereas the fuzzy mean algorithm gives an accuracy of 93.5 with the significance value 0.038 for accuracy and 0.073 for sensitivity. Conclusion: The outcome of the study confirms that the CNN-based model provides better results in enhancing the accuracy of disease identification in mango leaves.