Non-communicable diseases (NCD) are the leading cause of mortality and disability worldwide. A balanced diet is a cornerstone in the prevention of both the onset and the progression of NCD. The Mediterranean diet (MD) has demonstrated efficacy to decrease NCD risk among high risk and general populations. The MD is characterized by the consumption of specific food groups (vegetables, fruits, legumes, whole grain products, etc.), their respective proportion, as well as meal consumption pattern. The aim of MediPiatto is to develop a smartphone-based system using artificial intelligence (AI) for the automatic meal assessment with respect to its adherence to MD. The assessment will be performed on 3 levels: per meal, per day and per week. The nutritionists involved in the project will identify a list of food groups and estimate their respective serving portions to optimally characterize a meal's adherence/non-adherence to MD in the form of a questionnaire. Food images, gathered by Oviva, will be distributed to a number of annotators (crowdsourcing will be used) , which will be asked to fill out the questionnaire/meal image. Thus, an annotated food image database will be created and used to train the AI algorithms. The system will use as input a meal image taken by a smartphone's camera. The image will be analysed by a number of algorithms based on computer vision, AI and rule-based techniques. Once training is completed, for each image the questionnaire will be automatically answered by the system and an immediate high-level, qualitative feedback will be provided to the user with a traffic light colouring code. The daily intake of the selected food groups will be estimated and visualized to the user. Moreover, its adherence to MD will be automatically assessed using a sliding window of 7 days and already established scoring methodologies. The system after systematic technical validation will be integrated into Oviva's commercially successful diet management platform.