Implicit-Scale 3D Reconstruction for Multi-Food Volume Estimation from Monocular Images

TL;DR

This paper introduces a new benchmark dataset for 3D food reconstruction from monocular images, emphasizing implicit scale inference and geometric reasoning to improve portion estimation accuracy in realistic dining scenarios.

Contribution

The paper presents a novel benchmark dataset and framing of food volume estimation as an implicit-scale 3D reconstruction problem, advancing geometric methods over appearance-based approaches.

Findings

Geometry-based methods outperform vision-language models in accuracy.

Top methods achieve 0.21 MAPE in volume estimation.

Benchmark highlights challenges of scale ambiguity and occlusion in real-world scenes.

Abstract

We present Implicit-Scale 3D Reconstruction from Monocular Multi-Food Images, a benchmark dataset designed to advance geometry-based food portion estimation in realistic dining scenarios. Existing dietary assessment methods largely rely on single-image analysis or appearance-based inference, including recent vision-language models, which lack explicit geometric reasoning and are sensitive to scale ambiguity. This benchmark reframes food portion estimation as an implicit-scale 3D reconstruction problem under monocular observations. To reflect real-world conditions, explicit physical references and metric annotations are removed; instead, contextual objects such as plates and utensils are provided, requiring algorithms to infer scale from implicit cues and prior knowledge. The dataset emphasizes multi-food scenes with diverse object geometries, frequent occlusions, and complex spatial arrangements. The benchmark was adopted as a challenge at the MetaFood 2025 Workshop, where multiple teams proposed reconstruction-based solutions. Experimental results show that while strong vision--language baselines achieve competitive performance, geometry-based reconstruction methods provide both improved accuracy and greater robustness, with the top-performing approach achieving 0.21 MAPE in volume estimation and 5.7 L1 Chamfer Distance in geometric accuracy.