Abstract

Food allergy occurs when the body’s immune system responds strangely to a particular food. Allergic reactions are frequently minor, but they can still be dangerous. Food intolerance and food allergy are not the same things. Symptoms of food intolerance can be brought on by issues with the same substances, including lactose, being digested. However, there are no allergic reactions. A large amount of food must be consumed than is necessary for an allergy to be triggered before experiencing symptoms of food intolerance. Contrary to allergies, a food intolerance never poses a life-threatening risk.

Our project aims to create an Android-based (mobile-based) application that can identify British items accurately. The system will identify the allergens present in the ingredients used and will alert the user, it will also assist the user in knowing the calorie intake. The system is an extension for visually challenged people or people who require assistance to read and write. The major concern is the system configuration and training time as we are working with images, thus restricting our scope to eleven British items.

The methodology includes creating an Android/iOS application using flutter as a front end, classification of images using the CNN model, inclusion of speech functionality and checking for allergens, and determining calorie intake.

We have limitations in building a complete application. Millions of images are required per class to build a model capable of predicting appropriately. It's very difficult to achieve this as the existing system configuration is very low compared to the requirements and the system requires GPU. Our future work will include creating a complete application that can accurately predict the British food items displayed in a food festival or an event.

Computer Vision and Speech Recognition – Auto Recipe detection and alert system using AI