Dish Detection in Indian Food Platters: A Computational Framework for Diet Management

Mansi Goel, Shashank Dargar, Shounak Ghatak, Nidhi Verma, Pratik Chauhan, Anushka Gupta, Nikhila Vishnumolakala, Hareesh Amuru, Ekta Gambhir, Ronak Chhajed, Meenal Jain, Astha Jain, Samiksha Garg, Ganesh Bagler

Abstract: Diet is central to the epidemic of lifestyle disorders. Accurate and effortless diet logging is one of the significant bottlenecks for effective diet management and calorie restriction. Dish detection is a challenging problem in Indian platters due to a visually complex traditional food layout. We present a comparative analysis of deep-learning-based object detection models for the 61 most popular Indian dishes. Rooted in a meticulous compilation of 68,005 platter images with 134,814 manual dish annotations, we first compare ten architectures for multi-label classification to identify ResNet152 (mAP = 84.51%) as the best model. YOLOv8x (mAP = 87.70%) emerged as the best model architecture for dish detection among the eight deep-learning models implemented after a thorough performance evaluation. By comparing with the state-of-the-art model for the IndianFood10 dataset, we demonstrate the superior object detection performance of YOLOv8x for this subset and establish Resnet152 as the best architecture for multi-label classification. The models trained on such a rich dataset have diverse applications for diet logging, food recommendation systems, nutritional interventions, and mitigation of lifestyle disorders. The proposed computational framework is extendable to include staple dishes from across global cuisines.