Classification of Medicinal Wild Plants Using Radial Basis Function Neural Network with Least Mean Square

Wild plants are considered as plants whose growth interferes with other plants. However, in its development it turns out that wild plants contain ingredients for medicines. However, information and knowledge about wild plant species, especially plants with herbaceous vegetation, needs to be carried out with various technologies, including through digital image processing. This study aims to develop a classification model for medicinal plant species using color and texture extraction using the Radial Basis Function Neural Network (RBFNN) algorithm with a combination of the Least Mean Square (LMS) algorithm. Color feature extraction is obtained by calculating the RGB average value and texture feature extraction using a Gabor filter obtained from the mean, entropy, and variance parameters. The result of feature extraction becomes input for RBFNN with LMS. RBFNN has three layers that have feedforward properties that can assist in solving classification or pattern recognition problems. The LMS algorithm is used for learning or updating neural network weights. Based on the test results by calculating precision, recall, and accuracy, it gets a precision value of 92.50%, recall gets a value of 91.74, accuracy gets a value of 92.08. These results indicate that the developed model can classify wild plant species with medicinal properties well.