3rd Place Solution to Large-scale Fine-grained Food Recognition

2025-10-24

AI Summary

This work addresses large-scale fine-grained food recognition, aiming to enhance the accuracy of food image analysis in health applications. To tackle the challenges of small inter-class variance and large intra-class variation, we propose a composite loss function integrating ArcFace Loss and Circle Loss, coupled with meticulous hyperparameter optimization and a multi-model ensemble strategy—thereby significantly improving feature discriminability and classification robustness. Evaluated on the Kaggle LargeFineFoodAI-ICCV Workshop challenge, our method ranks third, demonstrating strong effectiveness and generalizability on real-world fine-grained food datasets. The core contributions lie in (1) the synergistic design of complementary metric learning losses and (2) a lightweight, efficient ensemble framework. This approach provides a reproducible, deployable solution for high-accuracy food recognition under resource-constrained conditions.

Technology Category

• Computer Vision: Large Vision Models
• Machine Learning: Ensemble Methods
• Natural Language Processing: Learning & Optimization for NLP

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Food analysis is becoming a hot topic in health area, in which fine-grained food recognition task plays an important role. In this paper, we describe the details of our solution to the LargeFineFoodAI-ICCV Workshop-Recognition challenge held on Kaggle. We find a proper combination of Arcface loss[1] and Circle loss[9] can bring improvement to the performance. With Arcface and the combined loss, model was trained with carefully tuned configurations and ensembled to get the final results. Our solution won the 3rd place in the competition.

Problem

• Developing fine-grained food recognition for health analysis
• Optimizing loss function combinations to improve model performance
• Winning third place in LargeFineFoodAI recognition challenge

Innovation

• Combined Arcface and Circle loss functions
• Trained model with carefully tuned configurations
• Used ensemble methods for final results