Publication

Abstract : This paper implements the deep learning system for detecting and estimating Indian food items and then estimating their nutritional content. The model is trained on Indian Food dataset with 17 Indian food classes.Data Augmentation techniques like mosaic, mixup, jitter were applied to improve generalization.The trained model achieved a strong mAP@50 score of 94 %.A nutritional estimation model was then introduced that uses calorie dataset and size factor to scale values based on the estimated food portion.This combined system is useful for smart diet monitoring,AI based kitchen applications and modern health monitoring tools.The results serve as a concrete proof that YOLOv11m can effectively detect diverse food items and provide nutritional insights in dynamic environment.