FastPano3D: Feed-Forward Indoor Panoramic 3D Reconstruction from a Single Image

2026-06-29

AI Summary

This work addresses the challenge of high-fidelity indoor 3D reconstruction from a single panoramic image, which is hindered by equirectangular projection distortion and non-uniform feature distribution. The authors propose the first end-to-end feedforward framework that directly generates renderable 3D Gaussian representations from a single panorama, without requiring multi-view supervision or test-time optimization. The method integrates a lightweight feature encoder, adaptive Gaussian sampling, and a point-cloud-guided refinement mechanism. Compared to the current state-of-the-art, it achieves a 156× speedup in inference, reduces model parameters by 50%, and delivers rendering quality on par with NeRF and 3D Gaussian Splatting.

Abstract

Recent advances in 3D scene reconstruction have highlighted the intricate trade-offs among rendering quality, inference efficiency, and data dependency. To address the challenge of rapidly reconstructing detailed 3D indoor scenes from minimal input, we introduce FastPano3D, an end-to-end framework that directly generates renderable 3D Gaussian representations from a single panoramic image. Unlike perspective-based methods, panoramic images inherently suffer from equirectangular projection distortions and spatially non-uniform feature distributions, making direct feed-forward Gaussian generation particularly challenging. In contrast to existing Gaussian Splatting based methods that rely on multi-view supervision or per-scene optimization, FastPano3D employs a lightweight feature encoder, adaptive Gaussian sampling, and a point-cloud-guided refinement strategy to achieve efficient and accurate scene generation without any test-time optimization. Our approach reconstructs high-fidelity 3D scenes within seconds, achieving up to 156 times faster inference than prior state-of-the-art methods such as Pano2Room, while using only half the parameters. Extensive experiments demonstrate that FastPano3D delivers rendering quality comparable to NeRF- and 3DGS-based reconstructions, establishing a new benchmark for rapid, single-view 3D scene inference.

Problem

• panoramic 3D reconstruction
• single-image 3D
• equirectangular distortion
• indoor scene reconstruction
• feed-forward 3D generation

Innovation

• FastPano3D
• 3D Gaussian Splatting
• single-view reconstruction
• panoramic image
• feed-forward inference