Image-Based Recognition of Herbal Plants and Flowers Using Convolutional Neural Networks for Innovative Learning

Abstract

In the current era of rapid technological advancements, there is a significant increase in the demand for technology applications, which are evolving at an astonishing pace. Consequently, obtaining appointments for minor illnesses at hospitals became increasingly difficult due to the overwhelming number of COVID-19 cases. However, being resilient as Filipinos are, the majority tried to find alternative ways to treat minor health concerns by using herbal plants. This study aims to determine the key functionalities required in a mobile application to facilitate innovative learning approaches and effectively enhance students' knowledge and understanding of herbal plants, including their applications and uses. Convolutional neural network is used to develop a real-time image-based recognition system for herbal plants and flowers and a dataset of herbal plants and flowers photos were used to train the system. The mobile application system's functionalities were assessed by learners using ISO 25010, and the findings reveal that the system can identify various herbal plants and flowers with high accuracy. It was concluded that the mobile application functionalities (3.92) are the highest consideration in designing the mobile application to achieve innovative learning when it comes to plants as compared to portability (3.86), usability (3.85), reliability (3.80), and maintainability (3.64).