A multi-feature selection approach for gender identification of handwriting based on kernel mutual information

Abstract

This paper presents a new flexible approach to predict the gender of the writers from their handwriting samples. Handwriting features like slant, curvature, line separation, chain code, character shapes, and more, can be extracted from different methods. Therefore, the multi-feature sets are irrelevant and redundant. The conflict of the features exists in the sets, which affects the accuracy of classification and the computing cost. This paper proposes an approach, named Kernel Mutual Information (KMI), that focuses on feature selection. The KMI approach can decrease redundancies and conflicts. In addition, it extracts an optimal subset of features from the writing samples produced by male and female writers. To ensure that KMI can apply the various features, this paper describes the handwriting segmentation and handwritten text recognition technology used. The classification is carried out using a Support Vector Machine (SVM) on two databases. The first database comes from the ICDAR 2013 competition on gender prediction, which provides the samples in both Arabic and English. The other database contains the Registration-Document-Form (RDF) database in Chinese. The proposed and compared methods were evaluated on both databases. Results from the methods highlight the importance of feature selection for gender prediction from handwriting.

Gender Determination Using Data Mining Methods Through Measurements of Ascender and Descender Parts of Letters

“…Cha and Srihari 20 Using ANNs, and the success rate was reported to be 70.2% SVM Bouadjenek et al 23 Reported that accuracy of approximately 70% was obtained regardless of the language in the gender determination study conducted on both Arabic (KHATT) and English (IAM) handwritten text clusters using histogram oriented gradient and gradient local binary pattern systems and the support vector machine (SVM) classifier One Class SVM Guerbai et al 24 Using one-class SVM (OC-SVM), the success rate with a single classifier was reported to be 62.49%, and it was 77.3% in the case of combined classifiers SVM Bi et al 41 In this study, an approach called Kernel Mutual Information (KMI) is proposed. It is 66.3% in the ICDAR 2013 database and 66.3% in the RDF.…”

Mass Spectrometry-Based Proteomic Discovery of Prognostic Biomarkers in Adrenal Cortical Carcinoma

“…To determine the important features, we employed three different algorithms: Relief [32], Information Gain [33], and analysis of variance (ANOVA) F-value [34] using the open-source software ORANGE (version 3.26; University of Ljubljana, Ljubljana, Slovenia). These feature selection models are well known for identifying features with good classification performance [35,36].…”

Author Profiling in Social Media with Multimodal Information

“…Mutual information measure provides a formal way to model the mutual information between the features and the classes (Bi et al, 2018). The mutual information M i (w)…”