Multi-Feature Aggregation in Diffusion Models for Enhanced Face Super-Resolution

TL;DR

This paper introduces a diffusion model-based algorithm that leverages multiple low-quality images and features to produce high-quality face super-resolution images, outperforming existing methods in recognition tasks.

Contribution

It is the first to combine multi-features with low-resolution images as conditioners in diffusion models for face super-resolution, improving reliability without explicit attribute input.

Findings

Achieved state-of-the-art facial recognition accuracy

Demonstrated robustness across multiple datasets

Outperformed existing super-resolution methods

Abstract

Super-resolution algorithms often struggle with images from surveillance environments due to adverse conditions such as unknown degradation, variations in pose, irregular illumination, and occlusions. However, acquiring multiple images, even of low quality, is possible with surveillance cameras. In this work, we develop an algorithm based on diffusion models that utilize a low-resolution image combined with features extracted from multiple low-quality images to generate a super-resolved image while minimizing distortions in the individual's identity. Unlike other algorithms, our approach recovers facial features without explicitly providing attribute information or without the need to calculate a gradient of a function during the reconstruction process. To the best of our knowledge, this is the first time multi-features combined with low-resolution images are used as conditioners to generate more reliable super-resolution images using stochastic differential equations. The FFHQ dataset was employed for training, resulting in state-of-the-art performance in facial recognition and verification metrics when evaluated on the CelebA and Quis-Campi datasets. Our code is publicly available at https://github.com/marcelowds/fasr