◇中华印刷科技学会第六届第二次会员大会专题演讲
考虑影像内容的跨媒体色域对映技术
Assessment of Digital Halftoing of Color Image
主讲者:陈鸿兴(Hung-Shing Chen)、阙家彬(Chia-pin Chuh)
![]() |
|
主讲者:陈鸿兴(Hung-Shing Chen) |
摘要:
随着计算机网络时代的来临,光电影像媒体亦跟随蓬勃发展,例如数字相机、彩色打印机、液晶屏幕、彩色扫描机、DVD等的高品质、低价格化迅速成长,促成影像媒体的急速彩色化。在这样以人眼视觉与计算机相互密切连接的开放环境中,负责不同媒体间色彩信号的连接及转换之色彩管理系统(Color Management System; 简称CMS)越来越扮演重要的角色。
装备独立色 (Device Independent Color)的概念逐渐成为影像输出入设备中CMS的基本设计思想。装备独立色的使用,使得不同媒体间的色彩信号可以相互正确地传达。但是,仅利用这种概念,仍无法解决不同媒体间色外貌 (color appearance)不一致的现象,特别是在屏幕影像和打印图像的比较上,色外貌的不一致尤为明显。当在同一照明、同一辉度的操作环境下,造成不同媒体间色外貌之差异,主要来自于两种不同媒体的色域大小差异,而要解决这个问题,必须靠CMS内的色域对映技术。一般而言,依照使用目的,色域对映可以分为色域压缩(gamut compression)与色域延展(gamut extension)两种技术。通常在银幕色域和打印机打印色域的相对关系中,因为屏幕色域较打印机打印色域来得大,所以大多数的跨媒体色域对映技术主要是指从屏幕色域到打印机打印色域之压缩技术。
利用数学算法解决跨媒体色域对映的方式称为色域对映演算(Gamut Mapping Algorithm,简称GMA)。到目前为止,大部分的GMA是建立在为设备和设备为基础的相互关系上,我们称这种方式为「设备主导」型GMA (Device-to-Device GMA)。但是,在设计这种GMA时,并没有将原来影像的色分布纳入考虑,所以容易导致压缩后的影像产生阶调损失、彩度低下的劣化结果。因此,有必要依照每种输入影像的色分布来设计能够考虑影像内容的「影像主导」型GMA (Image-to-Device GMA)。
另外,依照色空间次元性,GMA 可分为以下三种类型:(1)调整影像明度或彩度动态范围 (dynamic range) 的一次元对映方式。(2)固定色相角或色相范围,在明度-彩度平面上实行色域对映的二次元对映方式,(3)在三次元色空间中,实行色域对映的三次元对映方式。目前的色域对映技术大多采用一次元对映方式或二次元对映方式。但是,依照这样方式处理的结果,容易造成计算效率低落及阶调不连续的现象。所以,GMA有必要设计在三次元色空间中直接实行。
因此,我们认为结合以上之「影像主导」型GMA和三次元色域压缩技术,发展出能够对各个输入影像内容作因应考虑的「三次元影像主导型GMA」。将这种新型的色域对映技术应用到数字相机,彩色扫描机,CRT/LCD,彩色打印机等的各种影像输出入设备中,将有效建立完整且高画质的跨媒体色彩管理体系。
Abstract:
As the opto-electronic image I/O equipments (such as digital camera, color printer, LCD monitor, scanner etc.) are growing rapidly, it is necessary for users to require the image contents represented in true color. To make color signals within different devices transmit each other without errors and reproduce original color scenes faithfully, the development of cross-media Color Management System (CMS) is very important.
There are 2 kinds of color principles for image I/O devices, additive color mixing for luminous color and subtractive color mixing for non-luminous color. The former representation is display, and the latter one is color printer. Generally speaking, the calculations of CMY signals in color printer are more complex than RGB signals in display. When these 2 different types are connected in CMS of image devices, it is necessary to transform color signals by using Color Correction technologies.
Before the performance of Color Correction, Gamut Mapping Algorithm (GMA) will be applied to adjust device-1 colors (e.g., CRT monitor), which have bigger color gamut-size can be represented as the same color appearance as device-2 colors (e.g., inkjet printer) which have smaller color gamut. Therefore, GMA is one of the key technologies on the process of cross-media color reproduction. However, most of the current GMAs are based on the concept of Device-to-Device (D-D), which is negligent of the image color distributions. It tends to cause the over-compressed image colors. For solving this problem, the Image-to-Device (I-D) GMA considering the image contents is desired to develop.
In this paper, 2D basic model and 3D extension model of Image-to-Device g-compression GMA we developed are introduced. They can be well applied to the cross-media CMS.