Deep Learning and Augmented-Reality Glasses based Meat Cooking Support for Color Vision Disorder Compensation

People who suffer from color vision disorder (CVD) have difficulties recognizing colors and have limited capability to cook the meat, i.e., could not judge whether the meat is undercooked, or overcooked. Most CVD compensation approaches proposed in the existing studies focused on the problem of contrast loss experienced by CVD individuals, and these methods are not applicable to the meat-cooking task; while some studies proposed for this task may lose effect when the lighting condition changes. In this study, we propose a system for CVD support, which automatically determines the degree of cooking utilizing deep learning model and presents generated support information to CVD individuals using augmented-reality (AR) glasses. To enable the capability of the proposed system to perform the prediction, we create a meat-cooking dataset in this study. To confirm the effectiveness of the proposed method, quantitative and subjective evaluation experiments were conducted. The experimental results showed that the meat cooked with assistance of the proposed is better than those without the system.