Application scenario

Head wearing, no difference from ordinary glasses.

Monitoring scope

Guide a reasonable diet with a personal digital health library.

Automatic food recognition technology

A food image recognition method based on DCNN and transfer learning. According to the pre-trained DCNN model on the ImageNet image dataset, the network parameters are initialized, and then the fine-tuning training method is used to perform transfer learning on the self-built small-scale food image database set in order to obtain high-level attribute features of the food image. Finally, the high-level attribute features learned by DCNN are input to a linear support vector machine for classification of food images. The experiment proves that the accuracy of western-style restaurant recognition is 94.20%.



MR Mixed Reality Technology

Adopting video DTG perspective technology, high dynamic range image acquisition through binocular camera, real-time rendering through computer algorithms, new "realistic pictures" are obtained and presented from a large perspective. Through cloud computing, overlay some virtual images to achieve AR, or fully overlay virtual images to achieve VR.

Product Features
Diet management, intelligent assessment of food intake calories, carbohydrates, fats, proteins, etc.