TEXTRIX: Latent Attribute Grid for Native Texture Generation and Beyond

TL;DR

TEXTRIX introduces a native 3D attribute grid framework utilizing a Diffusion Transformer with sparse attention, enabling high-fidelity texture synthesis and precise 3D segmentation without multi-view fusion limitations.

Contribution

The paper presents a novel native 3D attribute generation method that improves texture quality and segmentation accuracy over existing multi-view fusion approaches.

Findings

State-of-the-art texture synthesis quality.

High-precision 3D part segmentation.

Seamless and detailed 3D textures.

Abstract

Prevailing 3D texture generation methods, which often rely on multi-view fusion, are frequently hindered by inter-view inconsistencies and incomplete coverage of complex surfaces, limiting the fidelity and completeness of the generated content. To overcome these challenges, we introduce TEXTRIX, a native 3D attribute generation framework for high-fidelity texture synthesis and downstream applications such as precise 3D part segmentation. Our approach constructs a latent 3D attribute grid and leverages a Diffusion Transformer equipped with sparse attention, enabling direct coloring of 3D models in volumetric space and fundamentally avoiding the limitations of multi-view fusion. Built upon this native representation, the framework naturally extends to high-precision 3D segmentation by training the same architecture to predict semantic attributes on the grid. Extensive experiments demonstrate state-of-the-art performance on both tasks, producing seamless, high-fidelity textures and accurate 3D part segmentation with precise boundaries.