<![CDATA[Food nutrition information plays an important role in helping individuals monitor their daily dietary intake. However, manually identifying food types is often impractical and prone to errors. Therefore, this study aims to develop an image-based food detection system capable of automatically displaying nutritional information. The method employed in this research is You Only Look Once version 8 Nano (YOLOv8n), which was selected due to its high inference speed and relatively low computational requirements. The dataset was collected from two sources, namely the internet and direct image acquisition, with a ratio of 2:1, resulting in a total of 2,000 images across 13 food classes. The experimental results show that the YOLOv8n model achieved an mAP@0.5 score of 89.3% on the test dataset. Several food classes obtained mAP values above 90%, while the remaining classes achieved scores ranging from 70% to 80%. In addition, the system successfully integrated the FatSecret API to retrieve and display nutritional information based on the detected food items, as well as to monitor users’ daily nutritional intake. These findings demonstrate that YOLOv8n can be effectively utilized for food type detection and can support the automated provision of nutritional information.]]>
Copyrights © 2026
