An intelligent dining scene experience

Alper Kanak, Ahmet Ozlu, Suleyman Olcay Polat, Ovgu Ozturk Ergun

Abstract: We present an interactive Mixed Reality (MR) system that merges an automatic image classification framework with 3D visualization of desired scenes for enhanced meal experience. The system is designed for Augmented Reality/Virtual Reality (AR/VR) glasses and provides an automatic scene selection scheme guided by actual food on the table and user's preliminary choices. First, Deep Neural Network (D-NN) based image classification framework recognizes food type from the image captured by AR/VR glasses. Later, 3D MR visualization framework combines real image of dining table with 3D generated scenes to display on the glasses in real-time.