Maintaining any diet occurs primarily through the fixation of when and what you eat. However, to monitor and record every single meal is not so simple. Experimental technology FitByte, which is being developed at Carnegie-Mellon (USA), is designed to solve this problem. Its essence lies in the use of multiple sensors that can be installed on any sunglasses manufacturer.

The base unit of the system is an infrared mechanism for fixing the movement, which is located on the front of the glasses. It detects a characteristic hand gestures related to eating and drinking. Such movements trigger the camera that captures the image of the consumed foods or beverages. Associating the received data, the system makes it possible to identify specific patterns of food behavior. At this stage of development the user still needs to determine what foods and beverages were photographed. Ultimately, however, the developers completely give this smart computer vision system.

Also FitByte there are six inertial sensors, which are located in touching glasses and nose. They fix associated with chewing jaw movements and vibration of the throat when swallowing.

All data is processed offline within the points, keeping record of what the user ate and drank, in what quantities and at what time. In the future the system can be added non-invasive sensors that will monitor the glucose level in blood and other physiological parameters, and also developed a special companion app. According to reports, the sales of the product will be around in three years.
Source — Carnegie Mellon University