Artificial Neural Network for image based Leaf Type Classification with Feature Extraction GLCM Method

One or more non-green leaf regions are mosaicked together to create variegated leaves. Early interest in the taxonomy of various leaf species was sparked by the various leaf folding patterns. Previous research tried to identify the primary disease kind that affects tomato and grape leaves. Multi-Crop Leaf Disease (MCLD) detection is done using a convolutional neural network (CNN). The created model more accurately categorizes sick leaves. In the hypothetical trial, the study's accuracy was 95.71% for tomatoes and 98.40% for grapes. The Artificial Neural Network (ANN) classification approach is suggested for this study because of the abundance of studies that have been conducted on various leaf types. This study used 5-step iteration optimization to calculate the number of folds utilizing loops from a minimum of 2 to a maximum of 10, and 21 characteristics were used. The dataset consisted of 275 different types of leaves. In this study, the accuracy was 92.8% at 4751 iterations, the number of folds at the eighth iteration, 100 neural training cycles, a 0.5 neural learning rate, and 0.6 momentum, and the RMSE was 0.264 with a standard deviation (+/-) of 5.73%.