Proceedings PaperUse of convex deficiencies for the recognition of hand-printed digits
|Format||Member Price||Non-Member Price|
Given the number and variety of methods offered for hand printed character recognition, it may very well be that there is no single method that can be called the `best.' Most algorithms either use a collection of properties of the set of pixels belonging to the object, or attempt to measure structural properties of higher level features (e.g., lines) which are composed of object pixels. Unfortunately the degree of variation seen in handprinted characters makes feature properties hard to measure accurately, and complicates the extraction of structural features. What is suggested here is the use of the background areas remaining after the character itself is removed.