Character Recognition Techniques and approaches: a literature review

Abstract

Researchers have carried out many approaches to recognizing the OCR through software-based and FPGA-based and its relationship with the ANNs and training these NNs by using the BP algorithm. The FPGA supports high speed to recognize the character because it works in parallel. It involves many logic circuits and can process at the same time (clock cycle). There have been problems with the established system because the size, shape, and style of supposedly identical characters might differ from person to person and even within the same person on rare occasions. The photograph is vulnerable to noise and can lose some or all of its visual clarity as a result.  There are strict guidelines that govern the way a character looks that the character must adhere to. The NN and its learning process could only be defined by heuristically deducing its rules from samples and the enormous dataset.