Character recognition for standard E-13B Magnetic Ink characters on cheques

摘要

The character recognition commences with scanning a character to obtain a waveform representative of the character. The waveform is then processed to identify the character. The processing commences by sampling the waveshape at a predetermined frequency to provide N samples. The samples are compared to a stored set of sample values from a reference character set. The closest match of the recorded samples to the sample set from a reference character identifies the character scanned. The closest match is determined by taking the difference between the amplitude of the scanned samples and of the reference samples.