SeeThrough3D: Occlusion Aware 3D Control in Text-to-Image Generation

TL;DR

SeeThrough3D introduces an occlusion-aware 3D scene representation for text-to-image generation, enabling realistic occlusions and precise camera control in multi-object scenes.

Contribution

The paper presents SeeThrough3D, a novel model that explicitly models occlusions using a translucent 3D scene representation and integrates it with pretrained text-to-image models.

Findings

Effective modeling of occlusions in 3D scene generation

Generalizes to unseen object categories

Enables precise 3D layout control with realistic occlusions

Abstract

We identify occlusion reasoning as a fundamental yet overlooked aspect for 3D layout-conditioned generation. It is essential for synthesizing partially occluded objects with depth-consistent geometry and scale. While existing methods can generate realistic scenes that follow input layouts, they often fail to model precise inter-object occlusions. We propose SeeThrough3D, a model for 3D layout conditioned generation that explicitly models occlusions. We introduce an occlusion-aware 3D scene representation (OSCR), where objects are depicted as translucent 3D boxes placed within a virtual environment and rendered from desired camera viewpoint. The transparency encodes hidden object regions, enabling the model to reason about occlusions, while the rendered viewpoint provides explicit camera control during generation. We condition a pretrained flow based text-to-image image generation model by introducing a set of visual tokens derived from our rendered 3D representation. Furthermore, we apply masked self-attention to accurately bind each object bounding box to its corresponding textual description, enabling accurate generation of multiple objects without object attribute mixing. To train the model, we construct a synthetic dataset with diverse multi-object scenes with strong inter-object occlusions. SeeThrough3D generalizes effectively to unseen object categories and enables precise 3D layout control with realistic occlusions and consistent camera control.