Abstract | ||
---|---|---|
This paper introduces the Masked Hamming-Distance (MHD) pattern recognition method which has originally been designed as general patternclassifier under the academic aspects of object-orientation and parallellism together with low space and time complexity. Two versions of the pattern-matcher (PATTMA) have been prepared. One is a strict sequential implementation of the recognition method, the other uses directly the parallelism of the associative processor AM3 which has also been developed at the department for technical computer sciences at J.W.Goethe-University as contribution to the PROMETHEUS project. PATTMA has been embedded into the character recognition-engine AQUIRE. AQUIRE is a heterogenous character recognition engine combining pure pattern-classification with stroke-evaluation to obtain best recognition results.The recognizer has been optimized to meet commercial requirements to support pen-based applications on personal computers with a powerful recognition-engine. The parallel implementation of AQUIRE has successfully been used within PROMETHEUS as recognizer for finger-printed guestures on a touch-panel to provide a comfortable user (driver) — interface to a car-co-pilot. |
Year | DOI | Venue |
---|---|---|
1993 | 10.1007/3-540-57233-3_78 | CAIP |
Keywords | Field | DocType |
handprinted characters,object-oriented pen-based recognizer,object oriented,hamming distance,pattern recognition,time complexity | Content-addressable memory,Object-oriented programming,Logical matrix,Character recognition,Pattern recognition,Intelligent character recognition,Computer science,Feature (machine learning),Artificial intelligence,Application programming interface,Executable | Conference |
ISBN | Citations | PageRank |
3-540-57233-3 | 1 | 0.52 |
References | Authors | |
1 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Bernd Klauer | 1 | 50 | 14.36 |
Klaus Waldschmidt | 2 | 122 | 30.92 |