A method for measuring dish size and providing nutritional information using image recognition and augmented reality based on artificial intelligence, which is performed by an apparatus, the method comprising the operations of:
- classifying and detecting at least one dish and at least one food contained in each dish from at least one image when the at least one image is received;
- predicting the size of each detected dish based on a learning model;
- predicting the volume of each detected food using the predicted size of the dish based on the learning model;
- and
- generating nutritional information using the predicted volume of each food and pre-stored food information,
- wherein any one among a first method and a second method is used when the size of each detected dish is predicted, the first method is a method of realizing and using the at least one image in augmented reality, and the second method is a method of detecting and using at least one bounding box from the at least one image.