| 摘要 |
<p>Preliminary classification of a text character (1, T, N, R) by examining a few of its PELs (picture element) rather than requiring repeated comparisons of the entire pattern with a large set of prototypes is achieved by following a path through a decision tree in a root-to-leaf node direction. Each tree node (2,4) specifies the location (7,14; 21,17) of a PEL in the character and the input pattern PEL value in turn determines the next branch (w, b), white or black to follow in the tree. The root node (2) identifies the first PEL to be inspected while the final leaf (1, 3, 5) indicates the class (I, T, N) selected. The library pattern and inputs are compared. Upon match (T=T), an index number is generated and the next pattern processed. Upon a mismatch (T#R) the input pattern (R) will be registered as a new prototype (R) in the library, the tree extended to include new nodes (6) and a new leaf (7) for the new pattern class, and the next pattern will be processed.</p> |
