Identification of Hoya Plant using Convolutional Neural Network (CNN) and Transfer Learning

Indonesia has the highest species diversity of the Hoya plant. The genus is increasingly popular due to the beauty of the flowers and leaves, and the number of hobbyists is still increasing. As the number of species and cultivar is increasing during cultivation, the identification of each species or cultivar then become problematic for people and hobbyist. An easy and quick identification system is urgently needed, especially the application embedded in the android or IOS system. This study aimed to build a Hoya identification model using the Convolutional Neural Network (CNN) to make it easier for Hoya determination. The resulting identification model identified Hoya well, with an accuracy of 90.476%.