Ultraflash cuts megapixel image synthesis time in half while improving quality
UltraFlash: Accelerating Megapixel Visual Synthesis
Computer Vision and Pattern Recognition
Summary
Creating very high-resolution images using AI models usually takes a lot of time because of repeated conversions between image and compressed forms. The authors found that this back-and-forth process is a major slowdown. They designed UltraFlash, which skips these repeated conversions by learning shortcuts and improves final image stitching. This makes the process about twice as fast while also making images look better.
What this means in practice
- •For visual effects studios: Generate ultra-high-resolution images faster by reducing bottlenecks in image decoding and resizing steps during AI-based synthesis.
- •For game developers: Create detailed game textures efficiently for large resolution displays by accelerating multi-stage image refinement pipelines.
Authors
Phuc Lai, Anh Nguyen, Phong Nguyen, Anh Tran
Abstract
Megapixel visual synthesis with latent diffusion models requires operating far beyond the resolutions at which these models are typically trained. Cascade refinement is a leading strategy for this setting: it starts from a base image at a native resolution and progressively upsamples it to the target resolution through multiple refinement stages. Each stage typically performs a pixel-space resolution transition followed by latent-space denoising. While prior acceleration efforts mainly focus on reducing denoising cost, we identify a complementary and underexplored bottleneck: the repeated movement of representations through the VAE. We call these operations pixel-space transitions. In cascade pipelines, intermediate latents are often decoded to RGB, resized, and re-encoded before refinement, while final latents are decoded patch by patch with dense overlap to suppress boundary artifacts. We propose UltraFlash, a transition-efficient approach that accelerates these VAE-heavy operations. UltraFlash has two components. Latent Hyperloop replaces intermediate RGB round trips with a learned latent-space shortcut that approximates the decode-resize-encode transformation without repeated VAE decoding and encoding. Sparse Patch-based Decoding accelerates final reconstruction by addressing patch seams caused by inconsistent GroupNorm statistics and reusing cached statistics across patches, enabling sparse VAE decoding with much smaller overlap. Integrated into a state-of-the-art cascade baseline, UltraFlash reduces 4K generation latency by approximately 2x while improving image quality. These results show that optimizing pixel-space transitions can improve both efficiency and fidelity, offering an effective route toward fast, high-quality megapixel visual synthesis.