Color Difference
The difference between two colors can only be quantified by reference to measured value. Quantifying color difference values is standard method used in quality control. For printed colors, it is conventional to report differences between the values of L*, a* and b* for the target and the sample.
Denoted by DL*, Da* and Db*, the difference distance in CIELAB space is denoted by DE, where
DE = [(DL*)2 + (Da*)2+ (Db*)2 ]1/2
Electronic Images - Computer graphic images can be stored in two types vector and raster.
Vector graphics - Graphics defined by mathematical descriptions of shapes, each of with may be assigned an outline with a given width and color. Also known as object-based graphics. Outline type such as PostScript or TrueType are actually examples of vector graphics.
Raster graphics - Raster graphics are defined in terms of pixels. In particular, raster graphics are defined by bitmaps which can be up to 32 bits in depth (for CMYK files). This makes possible the representation of millions of colors.
The simplest bit map is only one bit in depth, corresponding to only 2 levels of gray. Raster images can yield very large files. A 4x6 photographic scanned at 1200 dpi yields a 104MB RGB file and a 139MB CMYK file. Thus scaling and rotating are slow.
Scaling and rotating of vector graphics is fast since the operations are defined mathematically instead of pixel by pixel. Whether displayed or printed, vector graphics still needs to be rasterized. There are important differences, however. (A monitor is a 24-bit raster device while the imagesetter is a 2-bit raster device.) Nevertheless, for real time processing and storage, vector graphics is better for certain types of images.
Rasterization of lines or fonts may become aliased (or pixelized) when displayed or printed on a low resolution device. Vector and Raster objects may sometimes need to be converted into one another. Raster to vector conversion requires an autotrace routine that finds edges of objects.
Vector to raster conversion is done all the time in display and RIPs (Raster Image Processing). It is a simple round-off problem in terms of integer number of device pixels. However once converted conversion back to vector loses all mathematical information about its origin.