دانلود رایگان مقاله پیش بینی زیست سنجی نرم قوی از تجزیه و تحلیل دست خط آفلاین

Abstract

Currently, writer's soft-biometrics prediction is gaining an important role in various domains related to forensics and anonymous writing identification. The purpose of this work is to develop a robust prediction of the writer's gender, age range and handedness. First, three prediction systems using SVM classifier and different features, that are pixel density, pixel distribution and gradient local binary patterns, are proposed. Since each system performs differently to the others, a combination method that aggregates a robust prediction from individual systems, is proposed. This combination uses Fuzzy MIN and MAX rules to combine membership degrees derived from predictor outputs according to their performances, which are modeled by Fuzzy measures. Experiments are conducted on two Arabic and English public handwriting datasets. The comparison of individual predictors with the state of the art highlights the relevance of proposed features. Besides, the proposed Fuzzy MIN-MAX combination comfortably outperforms individual systems and classical combination rules. Relatively to Sugeno's Fuzzy Integral, it has similar computational complexity while performing better in most cases.

7. Discussion and comparison with the state of the art

The purpose of this work is to develop a combination paradigm to achieve robust gender, handedness and age range prediction from handwriting analysis. Recall that there are only few research works, which deal with this topic. From the results reported in Table 12, all previous works were carried out using private datasets, which does not favorite a fair quantitative comparison. Nevertheless, through methods observation, one can easily deduce the superiority of the proposed prediction systems. This is due to the incorporation of new topological and gradient features that allow local characterization of handwritten text. In addition, the Fuzzy MIN-MAX combination provides significant improvement compared to individual systems. • From all experiments, the main remark respective to softbiometrics prediction is that it is a language-independent task, since quite similar scores were obtained for Arabic and English corpuses. Unlike, IAM dataset in which, handwritten text is composed of detached characters. Arabic writing is semi-cursive, where a single word can be composed of several connectedcomponents. Also, Arabic language has its specific diacritical marking such as dumma (’), hamza (), or chadda (ω). Despite all these different properties, results of the blended corpus are typically in the same range as those of separated corpuses. • Another particularity of this work consists in performing feature generation by segmenting images into a uniform grid where features are calculated on each cell. This provides a local description of the image content. Experimental results highlighted the relationship between the gird size and the reliability of the feature characterization. Prediction accuracies obtained with individual systems, vary between 69% and 80%. The inspection along all datasets, reveals that the three SVM predictors provide satisfactory performance but there is no descriptor that allows the best discriminative power for SVM. According to the theory, such differences canpromote the accuracy improvement of a combination framework that contains not necessarily excellent classifiers, that disagree as much as possible on difficult cases [28].