Super-resolution Reconstruction of Single Image for Latent features

TL;DR

This paper introduces a novel latent feature-oriented diffusion model for single-image super-resolution, improving image quality, diversity, and sampling speed by encoding LR images and modeling complex distributions.

Contribution

It proposes a new LDDPM framework with a conditional encoder, normalized flow, and adversarial training to enhance super-resolution performance and efficiency.

Findings

Reconstructs more realistic HR images than existing methods.

Achieves better performance on multiple evaluation metrics.

Reduces sampling steps while maintaining image quality.

Abstract

Single-image super-resolution (SISR) typically focuses on restoring various degraded low-resolution (LR) images to a single high-resolution (HR) image. However, during SISR tasks, it is often challenging for models to simultaneously maintain high quality and rapid sampling while preserving diversity in details and texture features. This challenge can lead to issues such as model collapse, lack of rich details and texture features in the reconstructed HR images, and excessive time consumption for model sampling. To address these problems, this paper proposes a Latent Feature-oriented Diffusion Probability Model (LDDPM). First, we designed a conditional encoder capable of effectively encoding LR images, reducing the solution space for model image reconstruction and thereby improving the quality of the reconstructed images. We then employed a normalized flow and multimodal adversarial training, learning from complex multimodal distributions, to model the denoising distribution. Doing so boosts the generative modeling capabilities within a minimal number of sampling steps. Experimental comparisons of our proposed model with existing SISR methods on mainstream datasets demonstrate that our model reconstructs more realistic HR images and achieves better performance on multiple evaluation metrics, providing a fresh perspective for tackling SISR tasks.