3.2 . Controlled adaptive compression 
Controlled adaptive compression is a coding procedure adapting to different data 
sets without prior knowledge of the statistical properties. Two-dimensional data 
sets (images) are in a particular manner transformed into one-dimensional data 
sets (data sequences'). This sequence of d igital data is decomposed into partial se 
quences which are separately coded using an optimum code tor each sequence. 
The used code contains single value as well as length coding. The decomposition 
is preceded by a calculation whether a compression is feasible at all and if so 
which the most fitting code out of a set of available codes tor the particular sub 
sequence would be. IT is procedure avoids compression factors < i which may 
otherweise occur for complicated original data structures. 
This procedure has been applied to the classified image of Fig. ^ and achieved a 
compression factor of 6.84. A derailed description of the procedure is given in /2/. 
References: 
/1/ La Bonte, A.: Two-Dimesional Image Coding by Micro-Adaptive Picture-Se 
quencing 
Proceedings SP1 Volume 119: San Diego 1977 
/2/ Guenther, A.: Controlled adaptive compression - a redundancy reducing tech- 
ni q u e 
(del ivered fo r pu bl i ca tion) 
78