摘要
为了探索最新一代通用视频编码标准(versatile video coding, VVC)对中文文本屏幕内容图像(text screen content image, TSCI)感知质量的影响,设计图像主观观测实验并基于VVC混合编码框架原理研究了中文TSCI的VVC编码感知失真。构建中文文本屏幕内容图像数据库(Chinese text screen content image dataset, CT-SCID),设计图像主观观测实验,分析VVC引起的中文TSCI感知失真类型及其发展路径;结合VVC的混合编码框架原理,理论分析并实验验证影响VVC编码的中文TSCI感知失真程度的因素;总结当前代表性的屏幕内容图像质量评价方法在中文TSCI VVC编码感知失真评测上的性能表现。实验结果表明:字体大小和对比度是影响中文TSCI VVC编码感知质量的重要因素,且中文TSCI的字体越小、对比度越低时,图像的感知质量等级越低;当前代表性的屏幕内容图像质量评价方法均无法给出完全符合人眼感知特性的质量评价结果。研究对于后续开发适用于中文TSCI的感知质量评价方法、高效编码方法等具有指导意义。
To explore the influence of the state-of-the-art versatile video coding(VVC)on the perceptual quality of the Chinese text screen content image(TSCI),image subjective observation experiments were designed to study the perceptual distortion of Chinese TSCI using VVC based on the hybrid coding framework principles of VVC.A Chinese text screen content image dataset(CT-SCID)was constructed and image subjective observation experiments were designed to analyze the types and development paths of Chinese TSCI perceptual distortion caused by VVC.By combining the hybrid coding framework principles of VVC,factors that affect the degree of perceptual distortion of Chinese TSCI using VVC were theoretically analyzed and experimentally verified.The performance of representative screen content image quality evaluation methods for evaluating the perceptual distortion of Chinese TSCI using VVC was summarized.Experimental results show that font size and contrast are crucial factors affecting the perceptual quality of Chinese TSCI using VVC.A smaller font size and lower contrast of Chinese TSCI shall lead to a lower perceptual quality of the image.The existing representative screen content image quality evaluation methods are unable to provide quality evaluation results that fully conform to human visual perception characteristics.The research has some guiding significance for the subsequent development of perceptual quality evaluation methods and efficient coding methods applicable to Chinese TSCI.
作者
杨楷芳
晁学敏
蒙琴琴
公衍超
YANG Kaifang;CHAO Xuemin;MENG Qinqin;GONG Yanchao(School of Computer Science,Shaanxi Normal University,Xi’an 710119,China;Key Laboratory of Modern Teaching Technology,Ministry of Education,Xi’an 710062,China;Shaanxi Provincial Teaching Information Technology Engineering Laboratory,Xi’an 710119,China;School of Communications and Information Engineering,Xi’an University of Posts and Telecommunications,Xi’an 710121,China)
出处
《西安交通大学学报》
EI
CAS
CSCD
北大核心
2024年第4期18-31,共14页
Journal of Xi'an Jiaotong University
基金
国家自然科学基金资助项目(62277036)。
关键词
中文文本屏幕内容图像
通用视频编码
感知失真
笔画
Chinese text screen content image
versatile video coding
perceptual distortion
stroke