Narit HnoohomSumeth YuenyongMahidol University2019-08-232019-08-232018-06-081st International ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, ECTI-NCON 2018. (2018), 116-1192-s2.0-85050003696https://repository.li.mahidol.ac.th/handle/20.500.14594/45625© 2018 IEEE. This paper presents a prediction model for classifying Thai fast food images. The model uses a deep learning process that was trained on natural images (GoogLeNet dataset) and was fine-tuned to generate the predictive Thai fast food model. The researchers created a dataset, called the Thai Fast Food (TFF) dataset, which contained 3,960 images. The dataset was divided into eleven groups of food images comprised of omelet on rice, rice topped with stir-fried chicken and basil, barbecued red pork in sauce with rice, stewed pork leg on rice, Thai fried noodle, rice with curried chicken, steamed chicken with rice, shrimp-paste fried rice, fried noodle with pork in soy sauce and vegetables, wide rice noodles with vegetables and meat. The final group comprised dishes which are not members of the other ten groups listed (non-ten-types), but which exist among Thai fast food. The classification average accuracy on a separate test set shows that Thai fast food can be predicted at 88.33%.Mahidol UniversityComputer ScienceEngineeringConference PaperSCOPUS10.1109/ECTI-NCON.2018.8378293