Publisher Summary
This chapter reviews some mathematical models for graphics as data sources and the interplay between these models and coding schemes. Graphics include the following three classes of images: (1) images that are nominally two tone, (2) binary images derived from continuous-tone images, (3) and binary images created in some continuous-tone image coding schemes. A majority of the current efficient coding schemes for two-tone images are based directly or indirectly on the concept of transmitting contours or boundary points. The chapter also discusses the application of contour tracing, for which one needs to store the entire image at the transmitting and the receiving end. The most optimal and efficient way of coding uses the idea of pattern recognition. This is particularly appealing in the case of typewritten and printed matter. An important comparison one can make among the various coding schemes is how their compression factors depend on image resolution or sampling density.
[1]
Thomas S. Huang.
Coding of Two-Tone Images
,
1977,
IEEE Trans. Commun..
[2]
P. Wintz,et al.
A boundary finding algorithm and its applications
,
1975
.
[3]
P. Wintz.
Transform picture coding
,
1972
.
[4]
Thomas H. Morrin,et al.
A Black-White Representation of a Gray-Scale Picture
,
1974,
IEEE Transactions on Computers.
[5]
George Nagy,et al.
A Means for Achieving a High Degree of Compaction on Scan-Digitized Printed Text
,
1974,
IEEE Transactions on Computers.
[6]
Jack Capon,et al.
A probabilistic model for run-length coding of pictures
,
1959,
IRE Trans. Inf. Theory.