SECT - A coding technique for black/white graphics

A new coding technique called SECT for digitized two-tone or black/white pictures which is particularly applicable to polygonal objects composed of horizontal, vertical, and diagonal lines such as logical circuit patterns, characteristic font patterns, and mechanical drawings is described. While many previously proposed coding techniques attempt to encode the positions of transitive elements in a picture by using codes such as Huffman codes or Wyle codes which are constructed on the basis of picture statistics, selective element coding technique (SECT) focuses its attention on reducing the number of transitive elements in each picture independently of the statistics of the picture ensemble. The relation of the number of selective elements and objects in the picture is discussed. Furthermore the decoding algorithm to reproduce a picture from selective elements and its decodability are described.