A Comparison of Clustering Methods for Writer Identification and Verification

An effective method for writer identification and veri-fication is based on assuming that each writer acts as a stochastic generator of ink-trace fragments, or graphemes. The probability distribution of these simple shapes in a given handwriting sample is characteristic for the writer and is computed using a common codebook of graphemes obtained by clustering. In previous studies we used con-tours to encode the graphemes, in the current paper we ex-plore a complementary shape representation using normal-ized bitmaps. The most important aim of the current work is to compare three different clustering methods for gener-ating the grapheme codebook: k-means, Kohonen SOM 1D and 2D. Large scale computational experiments show that the proposed method is robust to the underlying shape rep-resentation used (whether contours or normalized bitmaps), to the size of codebook used (stable performance for sizes from 102 to 2:5 103) and to the clustering method used to generate the codebook (essentially the same performance was obtained for all three clustering methods). 1