Learning Perceptual Representations for Gaming NR-VQA with Multi-Task FR Signals

TL;DR

This paper introduces MTL-VQA, a multi-task learning framework that leverages full-reference metrics as supervisory signals to improve no-reference video quality assessment for gaming videos, especially when human-rated data is scarce.

Contribution

The novel approach uses multi-task learning with adaptive task weighting to learn perceptual features without human labels, enhancing NR-VQA performance on gaming videos.

Findings

MTL-VQA achieves competitive results with state-of-the-art methods.

It performs well in both supervised and self-supervised settings.

The framework effectively transfers learned features to NR-VQA tasks.

Abstract

No-reference video quality assessment (NR-VQA) for gaming videos is challenging due to limited human-rated datasets and unique content characteristics including fast motion, stylized graphics, and compression artifacts. We present MTL-VQA, a multi-task learning framework that uses full-reference metrics as supervisory signals to learn perceptually meaningful features without human labels for pretraining. By jointly optimizing multiple full-reference (FR) objectives with adaptive task weighting, our approach learns shared representations that transfer effectively to NR-VQA. Experiments on gaming video datasets show MTL-VQA achieves performance competitive with state-of-the-art NR-VQA methods across both MOS-supervised and label-efficient/self-supervised settings.