Estimating Indian Food Calorie with Convolution Neural Networks and Transfer Learning

Abstract

Listen

Translate article

The project's primary objective is to develop a robust and adaptable neural network model capable of recognizing the diverse range of Indian dishes from images. By harnessing the power of transfer learning, the model leverages pre-trained neural networks to extract relevant features, allowing for accurate identification of distinct dishes, even with limited training data. Endeavors addresses the vital need for precise and efficient estimation of calorie in Indian cuisine through the application of stateof-the-art transfer learning techniques in the realm of the food image recognition. Beyond food recognition, the project introduces a novel calorie estimation component. The neural network model, after identifying the dish in the image, utilizes its knowledge of the dish's ingredients and portion sizes to estimate calorie content. This integration of food recognizes and calorie content estimation serves a valuable tool for the individuals seeking to monitor their dietary intake and make informed nutritional choices, promoting a healthier lifestyle.