Identification of Philippine herbal medicine plant leaf using artificial neural network

• Main Authors: De Luna, Robert G., Baldovino, Renann G., Cotoco, Ezekiel A., De Ocampo, Anton Louise P., Valenzuela, Ira C., Culaba, Alvin B., Dadios, Elmer P.
• Format: text
• Published: Animo Repository 2017
• Subjects: Medicinal plants > Philippines—Identification; Image processing—Digital techniques; Neural networks (Computer science); Manufacturing
• Online Access: https://animorepository.dlsu.edu.ph/faculty_research/1914
• Institution: De La Salle University

The study described in this paper consists of a system that involves image processing techniques to extract relevant features related to leaf in conjunction with using artificial neural network in order to detect and identify some Philippine herbal plants. Real samples of twelve different herbal medicine plant leaves are collected where each leaf are isolated in single image. Several features are extracted using techniques in image processing. With the artificial neural network acting as autonomous brain network, the system can identify the species of the herbal medicine plant leaves being tested. The system can also provide information about the diseases the herbal plant can cure. For the training, a features dataset of 600 images coming from 50 images per herbal plant are used. With the aid of Python, a neural network model with optimized parameters are established producing 98.16 % identification for the whole dataset. To evaluate the actual performance of the system, a separate 72 sample images of herbal plants are tested with the neural network model implemented in MATLAB. Experimental results demonstrate a 98.61 % accuracy of herbal plant identification. © 2017 IEEE.