期刊名称：International Journal of Innovative Research in Computer and Communication Engineering
印刷版ISSN：2320-9798
电子版ISSN：2320-9801
出版年度：2016
卷号：4
期号：4
页码：7976
DOI：10.15680/IJIRCCE.2016.0404346
出版社：S&S Publications
摘要：Palm printidentification is an important personal identification and authentication method. Palm is explained as the inner surface of hand from the wrist to the root of the figures. In this paper we developthe accurate personal identification by combining the leftand right palm printimages. Palm printinghas gained the attention over several years from recent researches in multibiometrics. As the security system has very much important in several fields, it is very important to authenticate the users for any access. As many studies have been proposed but these researches did not explore the security issue in depth, so in this paper we established a framework in order to perform multibiometrics by combining left and right palm printimages. The framework which we implemented here requires pre-processing to remove the noisy portions and to enhance the image, Gaussian filter for enhancing the image. In this, the image is subjected to binarization with an average threshold value. The ROI portion was extracted based on bounding box calculation. After extracting the ROI part, the image was in gray scale form. The gray scale image was converted into binary image. This process of conversion is known as binarization. We used canny edge detection technique for detecting the edges of the image. Morphological operation is used for expanding and reducing the shape of the image. Here we used morphological Opening, which remove the unwanted pixels (small objects) which are present in the image. Gabor filter is used for palm image feature extraction. The extracted features of the left and the right palm image are combined and stored in the feature database. This process was repeated for all the images present in the input database. After this training process gets over, we choose a single testing image each from right and left palm printfolder respectively. If it was matched with the database, it results as genuine and if it was not matched, it results as fake. Thus recognition was achieved
关键词：Palm printing; Gabor filter; Canny edge detection; Normalization and ROI extraction
