TIMI: Training-Free Image-to-3D Multi-Instance Generation with Spatial Fidelity

Precise spatial fidelity in Image-to-3D multi-instance generation is critical for downstream real-world applications. Recent work attempts to address this by fine-tuning pre-trained Image-to-3D (I23D) models on multi-instance datasets, which incurs substantial training overhead and struggles to guarantee spatial fidelity. In fact, we observe that pre-trained I23D models already possess meaningful spatial priors, which remain underutilized as evidenced by instance entanglement issues. Motivated by this, we propose TIMI, a novel Training-free framework for Image-to-3D Multi-Instance generation that achieves high spatial fidelity. Specifically, we first introduce an Instance-aware Separation Guidance (ISG) module, which facilitates instance disentanglement during the early denoising stage. Next, to stabilize the guidance introduced by ISG, we devise a Spatial-stabilized Geometry-adaptive Update (SGU) module that promotes the preservation of the geometric characteristics of instances while maintaining their relative relationships. Extensive experiments demonstrate that our method yields better performance in terms of both global layout and distinct local instances compared to existing multi-instance methods, without requiring additional training and with faster inference speed.

Generating 3D scenes from a single image is useful for design, virtual reality, and digital content creation. However, when an image contains multiple objects, existing AI models often mix objects together or place them incorrectly. We propose TIMI, a training-free method that improves image-to-3D multi-instance generation. Instead of building a new model from scratch, TIMI guides the existing 3D generation model during its generation process. It encourages different instances to stay distinct while preserving their overall spatial arrangement. Experiments show that TIMI creates 3D scenes with better spatial layouts, clearer object separation, and efficient generation.