Efficient Encoder-Free Fourier-based 3D Large Multimodal Model

TL;DR

Fase3D introduces an encoder-free, Fourier-based 3D large multimodal model that efficiently processes large-scale point cloud data without heavy visual encoders, achieving competitive performance.

Contribution

The paper presents Fase3D, a novel encoder-free 3D LMM using Fourier transforms for efficient, scalable, and permutation-invariant scene understanding.

Findings

Fase3D matches encoder-based models in performance.

It significantly reduces computational cost and parameters.

The approach effectively models global context in large 3D scenes.

Abstract

Large Multimodal Models (LMMs) that process 3D data typically rely on heavy, pre-trained visual encoders to extract geometric features. While recent 2D LMMs have begun to eliminate such encoders for efficiency and scalability, extending this paradigm to 3D remains challenging due to the unordered and large-scale nature of point clouds. This leaves a critical unanswered question: How can we design an LMM that tokenizes unordered 3D data effectively and efficiently without a cumbersome encoder? We propose Fase3D, the first efficient encoder-free Fourier-based 3D scene LMM. Fase3D tackles the challenges of scalability and permutation invariance with a novel tokenizer that combines point cloud serialization and the Fast Fourier Transform (FFT) to approximate self-attention. This design enables an effective and computationally minimal architecture, built upon three key innovations: First, we represent large scenes compactly via structured superpoints. Second, our space-filling curve serialization followed by an FFT enables efficient global context modeling and graph-based token merging. Lastly, our Fourier-augmented LoRA adapters inject global frequency-aware interactions into the LLMs at a negligible cost. Fase3D achieves performance comparable to encoder-based 3D LMMs while being significantly more efficient in computation and parameters. Project website: https://tev-fbk.github.io/Fase3D.