| Many different mathematical techniques are used to solve pattern recognition problems. Some of them are considered efficient tools, but most of them are time-consuming and less reliable because of the influence of the technical difficulties in recognizing objects or characters. Several areas of technical difficulties still persist which deal with the resources and nature of the algorithms utilized. The difficulties are considered as the resource-inherent, computational, and methodological. The development of new algorithms is concerned with reduction of the difficulties, because the efficiency and robustness of a system is crucially dependent on the algorithms used in the recognition process.; The algorithms and their management techniques for the concurrent pattern recognitions are developed and implemented into computer systems. The pattern types and the method of concurrent processing for fast processing of the patterns are created first, and then the recognition tool and its system for the pattern classification are designed.; The proposed algorithm generates and optimizes pattern relationships. The pattern relationships are naturally derived from internal structures of patterns. The relationships of each pattern are used to define the pattern strength in a specified image. The pattern recognitions are performed concurrently based on their strengths. The algorithms and their management technique of the concurrent pattern recognition are developed and implemented. The technique is currently applied to Optical Character Recognition (OCR).; The OCR system under this technique produces dramatic results in terms of speed, reliability, and recognition rates. Even though the conventional approaches could not eliminate the technical difficulties, the new technique reduces the major difficulties of the system which is considerably faster and more reliable.; This dissertation includes an explanation of problems and difficulties of the old approaches, a proposed design and implementation, a review of the literature, a development of pattern types, a development of concurrent processing, a development of weight and entropy strengths, a description of the implementation, discussion of results, and a statement of conclusions. |
