# PortionNet: Distilling 3D Geometric Knowledge for Food Nutrition Estimation

arXiv: https://arxiv.org/abs/2512.22304 · HTML: https://mlsanity.com/paper/2512.22304/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 11 of its 24 references that are in the map.
- First posted on arXiv 26 Dec 2025.

## Builds on (most cited first, 11 of 11)

- [Distilling the Knowledge in a Neural Network](https://mlsanity.com/paper/1503.02531/) (2015)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [GradNorm: Gradient Normalization for Adaptive Loss Balancing in Deep Multitask Networks](https://mlsanity.com/paper/1711.02257/) (2017)
- [Cross Modal Distillation for Supervision Transfer](https://mlsanity.com/paper/1507.00448/) (2015)
- [Towards Learning Food Portion From Monocular Images With Cross-Domain Feature Adaptation](https://mlsanity.com/paper/2103.07562/) (2021)
- [Two-view 3D Reconstruction for Food Volume Estimation](https://mlsanity.com/paper/1701.03330/) (2017)
- [An End-to-end Food Portion Estimation Framework Based on Shape Reconstruction from Monocular Image](https://mlsanity.com/paper/2308.01810/) (2023)
- [Food Portion Estimation via 3D Object Scaling](https://mlsanity.com/paper/2404.12257/) (2024)
- [DPF-Nutrition: Food Nutrition Estimation via Depth Prediction and Fusion](https://mlsanity.com/paper/2310.11702/) (2023)
- [MFP3D: Monocular Food Portion Estimation Leveraging 3D Point Clouds](https://mlsanity.com/paper/2411.10492/) (2024)
- [Cross-Modal Distillation For Widely Differing Modalities](https://mlsanity.com/paper/2507.16296/) (2025)

## Abstract (arXiv)

Accurate food nutrition estimation from single images is challenging due to the loss of 3D information. While depth-based methods provide reliable geometry, they remain inaccessible on most smartphones because of depth-sensor requirements. To overcome this challenge, we propose PortionNet, a novel cross-modal knowledge distillation framework that learns geometric features from point clouds during training while requiring only RGB images at inference. Our approach employs a dual-mode training strategy where a lightweight adapter network mimics point cloud representations, enabling pseudo-3D reasoning without any specialized hardware requirements. PortionNet achieves state-of-the-art performance on MetaFood3D, outperforming all previous methods in both volume and energy estimation. Cross-dataset evaluation on SimpleFood45 further demonstrates strong generalization in energy estimation.