Point-MF: One-step Point Cloud Generation from a Single Image via Mean Flows

2026-04-27

AI Summary

This work addresses the challenge of reconstructing complete 3D point clouds from a single RGB image, which requires inferring occluded geometry. Existing diffusion models suffer from slow inference and high computational costs. To overcome these limitations, the authors propose Point-MF, the first framework to introduce Mean Flow to single-image point cloud reconstruction. By directly learning a mean velocity field in point cloud space, Point-MF enables one-step generation via a single forward pass without relying on a VAE latent space. The method integrates a Diffusion Transformer architecture with frozen DINOv3 image features and a lightweight token adapter, and introduces a Denoised Space Anchor loss to stabilize large-step generation, augmented by explicit time and interval conditioning. Evaluated on ShapeNet-R2N2 and Pix3D, Point-MF achieves millisecond-level inference and high-fidelity reconstructions, outperforming existing diffusion and feedforward models in both accuracy and speed.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Security and Privacy: Data transparency and provenance

Abstract

Single-image point cloud reconstruction must infer complete 3D geometry, including occluded parts, from a single RGB image. While diffusion-based reconstructors achieve high accuracy, they typically require many denoising iterations, resulting in slow and expensive inference. We propose Point-MF, a Mean-Flow-based framework for low-NFE single-image point cloud reconstruction that couples a Mean-Flow-compatible architecture with an auxiliary loss. Specifically, Point-MF operates directly in point-cloud space to learn the mean velocity field and enables one-step reconstruction with a single network function evaluation (1-NFE), without relying on VAE-based latent representations. To make Mean Flow effective under large interval jumps, Point-MF employs a Diffusion Transformer tailored to the Mean-Flow setting, conditioned on frozen DINOv3 image features via a lightweight token adapter and equipped with explicit interval/time conditioning. Moreover, we introduce Denoised Space Anchor, a set-distance auxiliary loss on the denoised-space estimate $x_θ$ induced by the predicted velocity field, to stabilize large-step generation and reduce outliers and density artifacts. On ShapeNet-R2N2 and Pix3D, Point-MF strikes a strong balance between reconstruction quality and inference speed compared to multi-step diffusion baselines and competitive feedforward models, while generating high-quality point clouds with millisecond-level latency.

Problem

• single-image 3D reconstruction
• point cloud generation
• occlusion completion
• fast inference
• 3D geometry recovery

Innovation

• Mean Flow
• one-step generation
• point cloud reconstruction
• Diffusion Transformer
• Denoised Space Anchor