Gimbal360: Differentiable Auto-Leveling for Canonicalized $360^\circ$ Panoramic Image Completion
This solves the challenge of generating seamless panoramic environments from in-the-wild images for applications like VR or 3D modeling, representing a novel method rather than an incremental improvement.
The paper tackles the problem of completing 360° panoramic images from unposed perspective inputs by addressing geometric and topological mismatches, achieving state-of-the-art performance in structurally consistent scene completion.
Diffusion models excel at 2D outpainting, but extending them to $360^\circ$ panoramic completion from unposed perspective images is challenging due to the geometric and topological mismatch between perspective projections and spherical panoramas. We present Gimbal360, a principled framework that explicitly bridges perspective observations and spherical panoramas. We introduce a Canonical Viewing Space that regularizes projective geometry and provides a consistent intermediate representation between the two domains. To anchor in-the-wild inputs to this space, we propose a Differentiable Auto-Leveling module that stabilizes feature orientation without requiring camera parameters at inference. Panoramic generation also introduces a topological challenge. Standard generative architectures assume a bounded Euclidean image plane, while Equirectangular Projection (ERP) panoramas exhibit intrinsic $S^1$ periodicity. Euclidean operations therefore break boundary continuity. We address this mismatch by enforcing topological equivariance in the latent space to preserve seamless periodic structure. To support this formulation, we introduce Horizon360, a curated large-scale dataset of gravity-aligned panoramic environments. Extensive experiments show that explicitly standardizing geometric and topological priors enables Gimbal360 to achieve state-of-the-art performance in structurally consistent $360^\circ$ scene completion.