As part of the MedDietAgent project, a food image dataset have been developed, the MedGRFood-Segmented database. The datasets contain food images of Greek and Mediterranean cuisine. The food images are divided into 11 food categories, based on the Greek Food Composition Dataset by the Hellenic Health Foundation. The food categories are the following: (i) Milk, dairy products or milk substitutes, (ii) Egg or egg products, (iii) Meat or meat products, (iv) Seafood or related products, (v) Fat or oil, (vi) Grain or grain products, (vii) Nut, seed or kernel products, (viii) Vegetable or vegetable products, (ix) Fruit or fruit products, (x) Sugar or sugar products, and (xi) Miscellaneous food products.
All images are annotated with respect to the food category, the exact food name, the cuisine and the weight of food.
The MedDietAgent food image dataset contains 9600 annotated images that belong to 96 food classes. MedGRFood-Segmented is the first annotated food image database of Mediterranean and Greek foods . Moreover, in MedDietAgent database have added annotated images from GlucoseML project. In total, we annotated and added 7000 more images that belongs to 70 classes, providing a complete food image dataset.
So, to date we have a total of 16600 food image labels belonging to 166 classes.
The food image database is available for research purposes after contacting with the Coordinator and Scientific Officer of the project, Professor Dimitrios I. Fotiadis, by filling in the adjacent contact form.