Unsupervised Curriculum Domain Adaptation for No-Reference Video Quality Assessment

During the last years, convolutional neural networks (C-NNs) have triumphed over video quality assessment (VQA) tasks. However, CNN-based approaches heavily rely on annotated data which are typically not available in VQA, leading to the difficulty of model generalization. Recent advances in domain adaptation technique makes it possible to adapt models trained on source data to unlabeled target data. However, due to the distortion diversity and content variation of the collected videos, the intrinsic subjectivity of VQA tasks hampers the adaptation performance. In this work, we propose a curriculum-style unsupervised domain adaptation to handle the cross-domain no-reference VQA problem. The proposed approach could be divided into two stages. In the first stage, we conduct an adaptation between source and target domains to predict the rating distribution for target samples, which can better reveal the subjective nature of VQA. From this adaptation, we split the data in target domain into confident and uncertain subdomains using the proposed uncertainty-based ranking function, through measuring their prediction confidences. In the second stage, by regarding samples in confident subdomain as the easy tasks in the curriculum, a fine-level adaptation is conducted between two subdomain-s to fine-tune the prediction model. Extensive experimental results on benchmark datasets highlight the superiority of the proposed method over the competing methods in both accuracy and speed. The source code is released at https://github.com/cpf0079/UCDA.

1. Introduction

Benefitting from the evolution of affordable and reliable consumer capture devices, and the tremendous popularity of social media platforms, recent years have witnessed an explosion of user-generated videos shared and streamed over the Internet [8]. Improving the efficiency of video coding, storage, and streaming over communication networks is a principle goal of video sharing and streaming platforms. One relevant research direction is the perceptual optimization of rate-distortion tradeoffs in video encoding and streaming, where distortion (or quality) is usually modeled using video quality assessment (VQA) algorithms that can predict human judgements of perceptual quality. This has motivated years of research on the topics of perceptual image and video quality assessment (IQA/VQA) [48], [36], [13], [22], [23], [12].