压缩人脸视频质量评估数据库（CFVQA）

压缩人脸视频质量评估数据库（CFVQA）是由香港城市大学计算机科学系的研究团队创建的，旨在系统理解人脸视频中的感知质量和多样化的压缩失真。该数据库包含3240个压缩人脸视频片段，这些片段来自135个内容多样的源视频，使用六种代表性的视频编解码器生成，包括两种基于混合编码框架的传统方法、两种端到端方法和两种生成方法。CFVQA的独特特点包括大规模、细粒度、内容多样性和跨压缩失真类型，使其成为现有图像质量评估（IQA）和视频质量评估（VQA）的可行和实用基准。该数据库的应用领域包括质量监控和优化，以及推动未来视频压缩算法的设计。Compressed Face Video Quality Assessment Database (CFVQA) was developed by a research team from the Department of Computer Science, City University of Hong Kong, with the goal of systematically understanding the perceptual quality and diverse compression distortions present in face videos. This database comprises 3,240 compressed face video clips sourced from 135 source videos with diverse content, and generated via six representative video codecs, including two traditional methods based on hybrid coding frameworks, two end-to-end methods, and two generative methods. The distinctive characteristics of CFVQA—large-scale, fine-grained, rich content diversity, and coverage of multiple compression distortion types—establish it as a feasible and practical benchmark for current image quality assessment (IQA) and video quality assessment (VQA) studies. Its application areas span quality monitoring and optimization, as well as advancing the design of next-generation video compression algorithms.

搜集汇总

背景与挑战

背景概述

CFVQA是一个由香港城市大学创建的人脸视频质量评估数据库，包含3240个压缩视频片段，源自135个多样源视频，使用六种编解码器生成。其特点是大规模、细粒度和内容多样性，适用于图像和视频质量评估基准，主要用于质量监控、优化和推动视频压缩算法设计。