摘要
计算机视觉算法的识别高度依赖图像质量,高动态范围图像拥有更大的亮度和色域范围,可以呈现更真实的自然场景和更逼真的视觉效果,从而被广泛的关注和应用。高动态范围图像采用浮点数存储,在传统设备上展现时会丢失大量信息,导致图像质量大幅度下降,常通过色调映射算法压缩动态范围,从而使二者动态范围相匹配,进行完整显示。文章对色调映射算法近些年的研究进行了综述,文章主要对传统色调映射与基于深度学习的色调映射的一些算法进行总结和探讨,分析了其中的优势和不足,并对色调映射算法进行简单总结与展望。
The recognition of computer vision algorithms is highly dependent on image quality.High dynamic range images have lar-ger brightness and color gamut range,which can present more realistic natural scenes and more realistic visual effects,so they are wid-ely concerned and applied.High dynamic range images are stored in floating-point numbers,which will lose a lot of information when displayed on traditional devices,resulting in a significant decrease in image quality.Tone mapping algorithm is often used to compress the dynamic range,so that the dynamic range of the two can match for complete display.This paper reviews the research of tone ma-pping algorithm in recent years.This paper mainly summarizes and discusses some algorithms of traditional tone mapping and deep learning-based tone mapping,analyzes their advantages and disadvantages,and briefly summarizes and prospects the tone mapping algorithm.
作者
金敏红
朱仲杰
娄晓涵
於至婧
卢豫哲
JIN Minhong;ZHU Zhongjie;LOU xiaohan;Yu zhijing;LU yuzhe(Zhejiang Wanli University,Ningbo 315100,China)
出处
《长江信息通信》
2023年第8期7-10,共4页
Changjiang Information & Communications
基金
浙江省自然科学基金
宁波市自然科学基金
浙江万里学院校级科研创新团队
浙江省属高校基本科研业务费项目成果
科技创新2025重大专项(宁波市重大科技任务攻关项目)。
