StereoSpace: Depth-Free Synthesis of Stereo Geometry via End-to-End Diffusion in a Canonical Space

2025-12-11

AI Summary

This work addresses monocular-to-stereoscopic image generation without explicit depth estimation or geometric warping. We propose an end-to-end diffusion-based paradigm that operates in a canonical rectified space, where view-conditioned embeddings directly model disparity distributions and occlusion-aware inpainting—enabling fully differentiable, parameterized training. To rigorously evaluate perceptual fidelity, we introduce a leakage-immune assessment protocol emphasizing downstream metrics: iSQoE (integrated Stereo Quality of Experience) and MEt3R (Multi-scale Edge-aware Temporal 3D Reconstruction error). Experiments demonstrate consistent superiority over warp-and-inpaint, latent-warping, and warped-conditioning baselines across both hierarchical and non-Lambertian scenes. Our method achieves state-of-the-art performance in disparity sharpness and geometric consistency, marking the first successful realization of high-fidelity stereo synthesis without depth prediction or explicit warping operations.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Security and Privacy: Large-scale security measurements
• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments

Abstract

We introduce StereoSpace, a diffusion-based framework for monocular-to-stereo synthesis that models geometry purely through viewpoint conditioning, without explicit depth or warping. A canonical rectified space and the conditioning guide the generator to infer correspondences and fill disocclusions end-to-end. To ensure fair and leakage-free evaluation, we introduce an end-to-end protocol that excludes any ground truth or proxy geometry estimates at test time. The protocol emphasizes metrics reflecting downstream relevance: iSQoE for perceptual comfort and MEt3R for geometric consistency. StereoSpace surpasses other methods from the warp&inpaint, latent-warping, and warped-conditioning categories, achieving sharp parallax and strong robustness on layered and non-Lambertian scenes. This establishes viewpoint-conditioned diffusion as a scalable, depth-free solution for stereo generation.

Problem

• Generates stereo images without depth data
• Ensures geometric consistency and perceptual comfort
• Outperforms existing methods in robustness and quality

Innovation

• Diffusion-based monocular-to-stereo synthesis without depth
• Canonical rectified space guides end-to-end correspondence inference
• Viewpoint-conditioned diffusion eliminates explicit warping and geometry estimates