High accuracy recognition techniques can provide useful information for the segmented handwritten and printed characters recognition systems. This research describes and analyses performances of a new statistical approach of features extraction for handwritten, printed and isolated numeral recognition.
The novel feature extraction approach is a combination of two statistical techniques, the first method is the ameliored transitions feature, where transitions between background and foreground pixels in the character image are calculated, and the second one is profile projection. Numeral recognition is carried out in this work through k nearest neighbors and fuzzy min max classifiers and shows a high recognition rate.