United States Photographic Industry: Color--Digital Colorization
Figure 1.--To colorize an old black and white photograph, a colorization artist has to spend a lot of time on it. Notice here the artist has not made the grass green or the sky blue. Everything except the boy's faces and one boy's hands has been left grey--showing how the process while digital was in no way automatic. A grey suit is not a bad guess for a suit, although the younger bnboy's tunic suit probably would not have been grey. And it is extrenely unlikely his bow would have been grey. The long stockings were probably black, although they could have been washed out. Put
With the perfection of color photography, interest turmed to ways colorizing black and white film. Film colorization is the y process that adds color to monochrome moving-picture images. It may be done as a special effect, to "modernize" black-and-white films, or to restore color films. The first examples date from the early 20th century, but colorization has become common with the advent of digital photography. This began with movies and then turned to old still photographs. This was possible through digital efforts. Computerized colorization began (1970s). The first technique was invented by Wilson Markle. These earliest attempts at colorization had soft contrast and were fairly pale, flat, washed-out color. They were easily identifuable as colorized images. The technology has improved steadily since the 1980s. Half-tone values were asigned colors. The problem is that no half-tone valur has unique color characteristics. Light blue and grey, for exaple, can not be distinguished. Thus there is no way of making colorization automatic. Some of the decisions were obvios, blue sky, green grass and foliage, flesh tones, etc. These are known as 'memory' colors. But many other colors have to be assigned by individual artists working with the process. Thus colorized images can look realistic, but the actual color choices are in actuality mostly guesses. he way the process works, a colotization artist begins by dividing the image into regions, and he then gives a color to each region. This technique is known as the segmentation method. It is certainly not automatiuc. It is extremely laborious and time-consuming, especially if there are no fully automatic algorithms to identify fuzzy or complex region boundaries. The area between a subject's hair and face is complicated to deal with. Colorization of movies with multiple frames probvides unique issues. They require motion compensation. This means tracking regions as movement occurs from frame to frame.
