Handwritten Numeral Recognition by Template Matching using Scilab

Handwritten character recognition is comparatively difficult, as different people have different handwriting styles. The aim of the present work is to recognize the numerals written by various persons and convert the recognized numerals into words. Template matching process is used to recognize the characters using normalised correlation method. Numerals 0-9 written by different persons with different style of handwriting are presented to the system. The presented character is used as image template which is compared with the search image. The proposed systems spot the detected character in the search image, show it into typed form and convert it into words. The accuracy of proposed system is 98.33 % and which can be further increased as system has flexibility to include more handwriting styles. The proposed system can be easily utilized in the application like OMR sheets where there is need to write the numerals values in figures as well as words. Scilab an open source software is used for recognition of these handwritten numerals.