Implementation high performance computing for medical plant classification

Medicinal plants are types of plants that are often used by the community because they have many benefits, such as to prevent or cure various diseases. Betel and binahong are medicinal plants that are widely used by Indonesian people. Part of the plant that is commonly used to classify plant species is the leaf. Convolutional Neural Network (CNN) is a very commonly method that used for image classification. This method produces the most significant accuracy in image recognition. Image processing using the Central Processing Unit (CPU) will require a long execution time. The Graphics Processing Unit (GPU) is one alternative to increase the speed of data processing. This research use 900 images of data with a comparison of training and testing data, namely 8:2. Based on global test results using GPU can increase data processing speed up to 55% compared by CPU without much affecting accuracy.