Thai Food Recognition Using Convolutional Neural Network with Dropout Technique

Visually recognizing Thai food is indeed a challenging task and tradition methods have achieved historically low classification. While deep learning operation allows users to identify Thai cuisine and visually distinguish Thai food, it often includes complex meanings. The deep learning approach has shown very satisfied results in such challenges. Therefore, this study focuses on presenting methods and techniques for noise reduction of Thai food images used in image recognition. A total of sixty thousand sample images were used in the experiment. Results revealed that the method offered highest reliable accuracy of 94.16%.