Comprehensive Comparative Analysis of Food Classification Using YOLOv8n, VGG19, and InceptionV3

Food classification plays a vital role in areas such as food recognition, nutritional tracking, and culinary exploration. This study evaluates the effectiveness of three deep learning models—YOLOv8n, VGG19, and InceptionV3—utilizing the In-dian Food Classification Dataset and Food-11. Metrics like accuracy, precision, recall, and F1 score indicate that InceptionV3 consistently delivers the highest accuracy, whereas YOLOv8n stands out for its real-time efficiency, making it well-suited for mobile or embedded systems. VGG19 provides a balanced performance but is limited by its larger model size and higher computational requirements. Statistical analysis (p ¡ 0.05) shows significant differences among the models, with confusion matrices revealing difficulties in classifying visually similar foods. The study suggests using InceptionV3 for tasks requiring precision, YOLOv8n for applications where speed is critical, and VGG19 for scenarios needing a balanced approach, while also proposing future research in cross-cultural food recognition, AI-driven food interaction, and personalized nutrition systems.