نوع مقاله : مقاله کامپیوتر
نویسندگان
1 گروه مهندسی کامپیوتر، دانشگاه صنعتی بیرجند، بیرجند، ایران
2 دانش آموخته کارشناسی ارشد، گروه مهندسی کامپیوتر، دانشگاه آزاد اسلامی، بیرجند
3 گروه مهندسی کامپیوتر، دانشگاه صنعتی بیرجند، بیرجند، ایران
4 دانشجوی پسادکتری، گروه مهندسی کامپیوتر، دانشگاه قطر
چکیده
کلیدواژهها
عنوان مقاله [English]
نویسندگان [English]
Recent advances in information technology and the need for more security have led to the rapid development of intelligent biometric identification systems. Recent studies have proven that handwriting of each person is unique and can be used as one of the authentication methods. There are many researches in the literature for writer identification on a specific language. Unfortunately, there are no necessary data sets for this purpose. In this paper, for the first time, a handwritten data set of 300 persons in both Persian and English languages was collected. The main goal of this paper is to provide a model to identify the writer independent of the language written in Persian and English. After pre-processing stage, each person's handwriting is converted into blocks of a certain size called a texture. Then, using these textures, the desired features are extracted. In order to extract these features, first a two-dimensional wavelet transform is applied to each image and then, using the new algorithm for calculating the fractal dimension, which is used for the first time in this field, the feature vector is obtained. Finally, MLP neural networks are utilized for classification step. The performance of the proposed method is evaluated in different scenarios.
کلیدواژهها [English]