Bilevel Pictures

Abstract

This chapter and the next two chapters deal with class 2 pictures where regions of fixed gray level or color are well defined. A major concern in the study of such images is the concept of shape, a term that is not easy to define quantitatively. One faces this problem when sampling an analog bilevel image. The size of the cells of the sampling grid must be small enough so that the shapes of regions of a given color remain unaltered in reconstructing the image. Sections 7.2, 7.4, and 7.6 deal with this and with other aspects of the digitization problem. Another set of problems in processing such images involves transforming them into a set of curves, or going back from a set of curves to plane regions. The tracing of the contour of a region is discussed in Section 7.5. Reconstructing the region from its contour, or filling, will be treated in Chapter 8. Instead of finding the contour, one may choose to thin a region to its skeleton, a matchstick type of figure whose curves or lines reflect the shape of the region. Thinning algorithms will be treated in Chapter 9.