The present invention relates to a system for monitoring diet using food recognition. The system is configured to capture a food image and utilize machine learning models to recognize the type of food being consumed, the ingredients, the food categories, and the nutrient of the dietary intake for each meal. Based on the classification of food categories and ingredients, the machine learning models provide a summary of a diet quality score for a particular meal. The system also provides real-time feedback and recommendation to the user to allow integration of behavioural change strategies for their log and future improvements in their daily diet intake. The system comprises an image recognition module, a cloud server, and a display module.