GoCARB. Smartphone app for carbohydrate calculation in type 1 diabetes

In recent years, new glucose monitoring devices and insulin delivery systems have significantly enhanced diabetes self-management. In parallel, recent advances in smartphone technology has permitted the introduction of a broad spectrum of applications to support diabetic patients in their everyday routine. The scope of the GoCARB is the automatic and near real-time estimation of a meal's carbohydrate content for individuals with type 1 diabetes based on computer vision and smartphone technologies. The user places a credit card-sized reference object next to the meal and acquires two images using a smartphone from different viewing angles. Then, a number of computer vision steps is executed. Initially, the food items on the plate are segmented and recognized while their 3D shape is reconstructed. Based on shape, segmentation results, and reference object, the volume of each item is estimated. Finally, the carbohydrate content is calculated by combining the food type with its volume, and using nutritional databases. For the design and development of GoCARB, a database of more than 5000 meal images has been created and used. Currently, the GoCARB system is under clinical evaluation, and therefore not yet accessible for public use.