Newest papers

Squaring-the-Sphere: Distortion-Free and Seamless Panorama Generation via CubeAtlas

Xuhui Zuo, Chenjie Cao, Tengfei Wang, Junta Wu, Yisu Zhang, Chunchao Guo · 2026 · arXiv (Cornell University)

DOI: 10.48550/arxiv.2610.31565 · cited by 0 works (OpenAlex) · open-access PDF

Spanish

La generación de panoramas de 360 grados requiere tanto la fidelidad de la perspectiva local como la continuidad esférica global. Las representaciones existentes se enfrentan a un compromiso: la proyección equidistante inferior (ERP) permite una modelización panorámica sin problemas de detalles, pero sufre de una grave distorsión y una brecha de dominio para generadores de imágenes pre-entrenados, mientras que los cubemaps conservan la geometría de la perspectiva orientada, pero a menudo producen caras inconsistentes y caras visibles. Presentamos la geometría de la perspectiva de los cubemaps, un marco de generación de imágenes panorámicas construido sobre una nueva representación del panorama interactivo CubeAtlas, que despliega todas las caras de cubemaps en una sola tela compacta. Al organizar las regiones polarizadas orientadas

Machine translation (Meta M2M100); the English original follows.

Abstract

Generating 360-degree panoramas requires both local perspective fidelity and global spherical continuity. Existing representations face a trade-off: equirectangular projection (ERP) enables seamless panorama modeling but suffers from severe distortion and a domain gap to pre-trained image generators, whereas cubemaps preserve perspective geometry but often produce inconsistent faces and visible seams. We present Squaring-the-Sphere, a panoramic image generation framework built on a novel representation CubeAtlas, which unfolds all cubemap faces into a single compact canvas. By arranging orientation-aligned, rotated polar regions alongside the equatorial faces, this representation preserves the perspective geometry of cubemaps while organizing the cross-face relationships required for seamless panorama assembly. It further enables a pretrained diffusion transformer to model the complete panorama jointly in a single pass, without altering the backbone architecture or positional encoding. While CubeAtlas has a token count comparable to ERP, its structured layout makes many dense attention interactions unnecessary. We therefore introduce a layout-aware sparse attention mask that preserves the interactions required for panorama continuity while removing geometrically irrelevant token pairs. As a result, CubeAtlas achieves better quality than ERP modeling with lower attention computation. Furthermore, we introduce a training-free super-resolution technique and a fusion scheme that enhance fine details in distortion-free perspective space before achieving the final ERP panorama. Experiments on various benchmarks demonstrate that our method improves geometric fidelity, cross-view consistency, and high-resolution visual quality over previous methods.

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Topics: Panorama, Computer science, Computer vision, Perspective (graphical), Artificial intelligence, Fidelity

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange