In this paper, we present a method of recognizing inclined, rotated characters. First we construct an eigen sub-space for each category using the covariance matrix which is calculated from a sufficient number of rotated characters. Next, we can obtain a locus by projecting their rotated characters onto the eigen sub-space and interpolating between their projected points. An unknown character is also projected onto the eigen sub-space of each category. Then, the verification is carried out by calculating the distance between the projected point of the unknown character and the locus. This method has the added advantage of obtaining the recognition result (category) and angle of inclination at the same time. In our experiment, we obtained 99.89 of recognition rate.
展开▼
