NAME AND RECIPE ESTIMATION OF THAI-DESSERTS BEYOND IMAGE TAGGING

Abstract

Thai-desserts or snacks (Thai: “ขนมไทย”) are included in Thai cuisine. The Food Republic – a well-known organization that gives some information about foods, drinks and snacks, has recommended the 7 Thai-desserts as Khanom Krok (Thai: ขนมครก), Coconut Ice-cream (Thai: ไอศกรีมกะทิสด), Fruity Luk Chub (Thai: ลูกชุบผลไม้), Woon Bai Toey (Thai: วุ้นใบเตย), Tub Tim Krob (Thai: ทับทิมกรอบ), Luem Gluen (Thai: ลืมกลืน) and Bua Loy (Thai: บัวลอย). However, the international outsiders do not dare to have these beautiful and delicious Thai-desserts because of the inconvenience of finding these Thai names and recipes. For example, some outsiders have diabetes and cannot have too much sugar or insulin. This paper combines the “Artificial Intelligence” and “Computer Vision” to build a computer model for tagging the name and recipe of unknown image. Our computer model was done on Convolution Neural Networks (CNN) that provided more than 85% of the accuracy.