PixelBoost: Leveraging Brownian Motion for Realistic-Image Super-Resolution

TL;DR

PixelBoost introduces a stochastic diffusion model leveraging Brownian motion to enhance the realism and efficiency of image super-resolution, focusing on texture and edge detail preservation.

Contribution

The paper presents a novel diffusion model that incorporates Brownian motion and stochasticity to improve super-resolution quality and training efficiency.

Findings

Superior LPIPS, PSNR, SSIM scores compared to existing methods

Enhanced edge reconstruction and texture detail

Faster inference with sigmoidal noise sequencing

Abstract

Diffusion-model-based image super-resolution techniques often face a trade-off between realistic image generation and computational efficiency. This issue is exacerbated when inference times by decreasing sampling steps, resulting in less realistic and hazy images. To overcome this challenge, we introduce a novel diffusion model named PixelBoost that underscores the significance of embracing the stochastic nature of Brownian motion in advancing image super-resolution, resulting in a high degree of realism, particularly focusing on texture and edge definitions. By integrating controlled stochasticity into the training regimen, our proposed model avoids convergence to local optima, effectively capturing and reproducing the inherent uncertainty of image textures and patterns. Our proposed model demonstrates superior objective results in terms of learned perceptual image patch similarity (LPIPS), lightness order error (LOE), peak signal-to-noise ratio(PSNR), structural similarity index measure (SSIM), as well as visual quality. To determine the edge enhancement, we evaluated the gradient magnitude and pixel value, and our proposed model exhibited a better edge reconstruction capability. Additionally, our model demonstrates adaptive learning capabilities by effectively adjusting to Brownian noise patterns and introduces a sigmoidal noise sequencing method that simplifies training, resulting in faster inference speeds.