DiT360: High-Fidelity Panoramic Image Generation via Hybrid Training

Haoran Feng, Dizhe Zhang, Xiangtai Li, Bo Du, Lu Qi

Keywords: image generation, panoramic image generation, hybrid training

TL;DR: A framework for high-resolution and photorealistic panoramic image generation with hybrid training.

Abstract:

In this work, we propose \textit{DiT360}, a DiT-based framework that performs hybrid training on perspective and panoramic data for panoramic image generation.
For the issues of maintaining geometric fidelity and photorealism in generation quality, we attribute the main reason to the lack of large-scale, high-quality, real-world panoramic data, where such a data-centric view differs from prior methods that focus on model design.
Basically, \textit{DiT360} has several key modules for inter-domain transformation and intra-domain augmentation, applied at both the pre-VAE image level and the post-VAE token level.
At the image level, we incorporate cross-domain knowledge through perspective image guidance and panoramic refinement, which enhance perceptual quality while regularizing diversity and photorealism.
At the token level, hybrid supervision is applied across multiple modules, which include circular padding for boundary continuity, yaw loss for rotational robustness, and cube loss for distortion awareness.
Extensive experiments on text-to-panorama, inpainting, and outpainting tasks demonstrate that our method achieves better boundary consistency and image fidelity across eleven quantitative metrics.
Our code, trained models, and curated data will be available.

Primary Area: generative models

Submission Number: 2030