Affine Invariant Recognition of Characters by Progressive Pruning

Abstract

There are many problems to realize camera-based character recognition. One of the problems is that characters in scenes are often distorted by geometric transformations such as affine distortions. Although some methods that remove the affine distortions have been proposed, they cannot remove a rotation transformation of a character. Thus a skew angle of a character has to be determined by examining all the possible angles. However, this consumes quite a bit of time. In this paper, in order to reduce the processing time for an affine invariant recognition, we propose a set of affine invariant features and a new recognition scheme called "progressive pruning." The progressive pruning gradually prunes less feasible categories and skew angles using multiple classifiers. We confirmed the progressive pruning with the affine invariant features reduced the processing time at least less than half without decreasing the recognition rate.

Histogram of Radon transform with angle correlation matrix for distortion invariant shape descriptor

“…In the case of invariant methods for any distortions, a method using ellipses was proposed to normalize such distortions [16,17]. This method estimates ellipses which approximates shapes.…”

Layout-free dewarping of planar document images

“…The recognition method for affinely distorted characters 13 is detailed. The principal of estimating the affine transformation matrix is illustrated in Fig.…”

“…For the sake of that, we estimate an affine transformation matrix by employing a recognition method for affinely distorted characters. 13 The method is based on the TW-values and an affine normalization as detailed later. One may think that if recognition of affinely distorted characters is possible, we do not need further rectification.…”

“…This problem is solved by the TW-value which is an improvement of DE-parameter proposed in. 13 The TW-value is calculated using two ellipses that are similar to the A-ellipse. Some examples are shown in Fig.…”