Adapting Pretrained Networks for Image Quality Assessment on High Dynamic Range Displays

TL;DR

This paper proposes methods to adapt pre-trained neural networks for high-dynamic-range image quality assessment, addressing the challenge of limited HDR data and improving performance over existing approaches.

Contribution

It introduces a domain adaptation approach that fine-tunes SDR-trained networks for HDR IQA, achieving better accuracy and faster convergence.

Findings

Models outperform previous baselines on HDR IQA datasets.

The adapted models converge more quickly than training from scratch.

The approach reliably generalizes to HDR content.

Abstract

Conventional image quality metrics (IQMs), such as PSNR and SSIM, are designed for perceptually uniform gamma-encoded pixel values and cannot be directly applied to perceptually non-uniform linear high-dynamic-range (HDR) colors. Similarly, most of the available datasets consist of standard-dynamic-range (SDR) images collected in standard and possibly uncontrolled viewing conditions. Popular pre-trained neural networks are likewise intended for SDR inputs, restricting their direct application to HDR content. On the other hand, training HDR models from scratch is challenging due to limited available HDR data. In this work, we explore more effective approaches for training deep learning-based models for image quality assessment (IQA) on HDR data. We leverage networks pre-trained on SDR data (source domain) and re-target these models to HDR (target domain) with additional fine-tuning and domain adaptation. We validate our methods on the available HDR IQA datasets, demonstrating that models trained with our combined recipe outperform previous baselines, converge much quicker, and reliably generalize to HDR inputs.