Off-Line Text-Independent Arabic Writer Identification using Contour-Based Features

In this paper, we show how the combination and cooperation of six feature vectors computed from the minimum-perimeter polygon (MPP) contours of Arabic words can lead to very interesting results for off-line text-independent Arabic writer identification. These feature vectors are in the form of probability distribution functions (PDFs), and are based on the length, direction, angle and curvature measurements. First, the aforementioned features were extracted and tested separately using 82 writers from the IFN/ENIT database. Then, they were combined in different ways. Finally, the combination which led to the best identification rates was retained. Indeed, promising results were obtained using a set of common distance metrics and the Borda ranking algorithm for classification. More specifically, we obtained 90.2% identification rate for Top1, and 97.5% for Top10.