GPU Based Image Classification using Convolutional Neural Network Chicken Dishes Classification

摘要

The internet makes human life easier, one of them is in orderingfood. By only using smartphones and opening food delivery features,in the application, consumers can order food from restaurants thatcollaborate with the e-commerce. The case study that we did wasclassifying the types of chicken dishes namely “ayam geprek”,“ayam goreng”, and “ayam bakar” using CNN method. CNNmethod is a popular method for image classification. This research isdoing image classification using CNN method based on GPU. Fromseveral models built, obtained the best model with an accuracy of99% and training speed about 233 seconds. While for comparison ofCNN method based on CPU and GPU obtained the conclusion thatthe most rapid training process using GPU.