Feedforward 3D Editing via Text-Steerable Image-to-3D

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Ma, Ziqi, Chen, Hongqiao, Yue, Yisong, Gkioxari, Georgia
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917147745189888
author Ma, Ziqi
Chen, Hongqiao
Yue, Yisong
Gkioxari, Georgia
author_facet Ma, Ziqi
Chen, Hongqiao
Yue, Yisong
Gkioxari, Georgia
contents Recent progress in image-to-3D has opened up immense possibilities for design, AR/VR, and robotics. However, to use AI-generated 3D assets in real applications, a critical requirement is the capability to edit them easily. We present a feedforward method, Steer3D, to add text steerability to image-to-3D models, which enables editing of generated 3D assets with language. Our approach is inspired by ControlNet, which we adapt to image-to-3D generation to enable text steering directly in a forward pass. We build a scalable data engine for automatic data generation, and develop a two-stage training recipe based on flow-matching training and Direct Preference Optimization (DPO). Compared to competing methods, Steer3D more faithfully follows the language instruction and maintains better consistency with the original 3D asset, while being 2.4x to 28.5x faster. Steer3D demonstrates that it is possible to add a new modality (text) to steer the generation of pretrained image-to-3D generative models with 100k data. Project website: https://glab-caltech.github.io/steer3d/
format Preprint
id arxiv_https___arxiv_org_abs_2512_13678
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Feedforward 3D Editing via Text-Steerable Image-to-3D
Ma, Ziqi
Chen, Hongqiao
Yue, Yisong
Gkioxari, Georgia
Computer Vision and Pattern Recognition
Artificial Intelligence
Recent progress in image-to-3D has opened up immense possibilities for design, AR/VR, and robotics. However, to use AI-generated 3D assets in real applications, a critical requirement is the capability to edit them easily. We present a feedforward method, Steer3D, to add text steerability to image-to-3D models, which enables editing of generated 3D assets with language. Our approach is inspired by ControlNet, which we adapt to image-to-3D generation to enable text steering directly in a forward pass. We build a scalable data engine for automatic data generation, and develop a two-stage training recipe based on flow-matching training and Direct Preference Optimization (DPO). Compared to competing methods, Steer3D more faithfully follows the language instruction and maintains better consistency with the original 3D asset, while being 2.4x to 28.5x faster. Steer3D demonstrates that it is possible to add a new modality (text) to steer the generation of pretrained image-to-3D generative models with 100k data. Project website: https://glab-caltech.github.io/steer3d/
title Feedforward 3D Editing via Text-Steerable Image-to-3D
topic Computer Vision and Pattern Recognition
Artificial Intelligence
url https://arxiv.org/abs/2512.13678