MFM-point: Multi-scale Flow Matching for Point Cloud Generation

2025-11-25

AI Summary

Point-based methods for point cloud generation suffer from poor scalability and low generation quality. To address this, we propose MFM-Point, a multi-scale flow matching framework that adopts a coarse-to-fine generation paradigm. It introduces a geometrically structure-preserving multi-scale sampling strategy to align distributions and enforce structural consistency across resolutions for unordered point clouds. Integrated with hierarchical down-sampling/up-sampling and flow matching, MFM-Point establishes an end-to-end, representation-free multi-scale generation pipeline. Crucially, it achieves these improvements without increasing training or inference overhead. Extensive experiments demonstrate that MFM-Point significantly enhances generation fidelity, achieving state-of-the-art performance among point-based methods on both multi-class and high-resolution point cloud benchmarks—and matching the quality of advanced representation-based approaches.

Technology Category

• Machine Learning: Multimodal Learning
• Natural Language Processing: Generation
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

In recent years, point cloud generation has gained significant attention in 3D generative modeling. Among existing approaches, point-based methods directly generate point clouds without relying on other representations such as latent features, meshes, or voxels. These methods offer low training cost and algorithmic simplicity, but often underperform compared to representation-based approaches. In this paper, we propose MFM-Point, a multi-scale Flow Matching framework for point cloud generation that substantially improves the scalability and performance of point-based methods while preserving their simplicity and efficiency. Our multi-scale generation algorithm adopts a coarse-to-fine generation paradigm, enhancing generation quality and scalability without incurring additional training or inference overhead. A key challenge in developing such a multi-scale framework lies in preserving the geometric structure of unordered point clouds while ensuring smooth and consistent distributional transitions across resolutions. To address this, we introduce a structured downsampling and upsampling strategy that preserves geometry and maintains alignment between coarse and fine resolutions. Our experimental results demonstrate that MFM-Point achieves best-in-class performance among point-based methods and challenges the best representation-based methods. In particular, MFM-point demonstrates strong results in multi-category and high-resolution generation tasks.

Problem

• Improves scalability and performance of direct point cloud generation methods
• Addresses geometric structure preservation across multi-scale resolutions
• Enhances generation quality without increasing training or inference costs

Innovation

• Multi-scale Flow Matching for point clouds
• Coarse-to-fine generation without extra overhead
• Structured downsampling and upsampling preserves geometry