ABSTRACT
Shapes in two-dimensional digitized images may be interpreted in a wide variety of ways. In some applications, they are projections of three-dimensional objects, whereas in others they may be sections through them. This latter case includes elevation profiles across surfaces, as can be provided by devices such as scanning profilometers. Numerical values based on measurements are not the only ways to describe shape. When a name can be applied, for instance, when a feature can be recognized as belonging to a known class, that name carries with it a presumed description of shape. Subsequent developments created fonts with a more “normal” appearance, and software has been developed that can read most standard fonts, both with and without serifs, and can adapt to different printing sizes. There is several commercial optical character recognition (OCR) software packages designed for use with scanners. OCR software may also use the topology of letters.
