SkeleGuide: Explicit Skeleton Reasoning for Context-Aware Human-in-Place Image Synthesis

TL;DR

SkeleGuide introduces explicit skeletal reasoning into human image synthesis, significantly improving structural accuracy and realism by learning internal pose representations and enabling user control.

Contribution

The paper presents SkeleGuide, a novel framework that explicitly models skeletal structure for improved human image synthesis, including a pose decoding module for user editing.

Findings

Outperforms existing models in image quality and structural plausibility.

Learned internal pose acts as a strong structural prior.

Enables fine-grained user control over generated images.

Abstract

Generating realistic and structurally plausible human images into existing scenes remains a significant challenge for current generative models, which often produce artifacts like distorted limbs and unnatural poses. We attribute this systemic failure to an inability to perform explicit reasoning over human skeletal structure. To address this, we introduce SkeleGuide, a novel framework built upon explicit skeletal reasoning. Through joint training of its reasoning and rendering stages, SkeleGuide learns to produce an internal pose that acts as a strong structural prior, guiding the synthesis towards high structural integrity. For fine-grained user control, we introduce PoseInverter, a module that decodes this internal latent pose into an explicit and editable format. Extensive experiments demonstrate that SkeleGuide significantly outperforms both specialized and general-purpose models in generating high-fidelity, contextually-aware human images. Our work provides compelling evidence that explicitly modeling skeletal structure is a fundamental step towards robust and plausible human image synthesis.