Nutrition Facts Recognition from Food Labels Images

Abstract

In this project we create a diet log application for the consumers based on the their calories intake from different food items. This way the consumer can maintain a regulatory diet and manage accordingly about his/her calories consumption for a day based on the limit set by the consumer itself. This project is implemented in different phases. The first phase includes the image pre-processing which includes median filtering, adaptive thresholding, histogram equalization and segmentation. The second phase includes the optical character recognition on the segmented data. Finally the total calories intake from each food items is added and an alert message is generated to notify the consumer if exceeded day’s limit.