A Feature Selection and Extraction Method for Uyghur Handwriting-Based Writer identification

Multi-Stage Based Feature Extraction Methods for Uyghur Handwriting Based Writer Identification

“…So these reports about writer identifications are mostly based on Latin handwriting (Plamondon et al, 1989;Said et al, 2000;Schomaker & Bulacu, 2004;Srihari et al, 2002), and Chinese handwriting (He et al 2007(He et al , 2008Li et al, 2009), Arabic handwritings (Al-Dmour & Zitar, 2007, even Persian handwritings (Helli & Moghaddam, 2010). However, there are only 4 reports about Uyghur handwriting based writer identification, in which two of them are our previous research (Ubul et al, 2008(Ubul et al, , 2009 indicated to using Gabor filter and Genetic algorithm (GA), and Gabor filter plus PCA and ICA methods for feature extraction, and get the 92.5% identification rate for 55 different people. Raxidin (2010) also used Gabor filter for feature extraction and achieved an accuracy rate of 79.8%.…”