| Summary: | Monitoring and managing dietary intake plays a vital role in maintaining a healthy lifestyle. However, accurately tracking food consumption and estimating calorie intake can be challenging. This paper presents a deep learning-based approach of You Only Look Once version 4 (YOLOv4) model for food detection and calorie counting. Leveraging the power of deep neural networks, the proposed study automatically detects and classifies food items from images and provides real-time estimation of their calorie content. Seven classes of food which are fried noodles, fried rice, kaya toast, nasi lemak, roti canai, fried chicken, and fried egg were addressed, by training the model on a diverse and well-annotated food dataset. We also tackle the issue of calorie estimation. Experimental evaluations of the proposed YOLOv4 for food detection demonstrates 96.07% of accuracy. Thus, it could be deduced that the proposed deep learning-based food detection and calorie counter have the potential to significantly improve dietary monitoring and contribute to the promotion of healthier eating habits. © 2023 IEEE.
|
|---|