A new method for sclera vessel recognition using OLBP

Publication Type:
Conference Proceeding
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2013, 8232 LNCS pp. 370 - 377
Issue Date:
Filename Description Size
10.1007%2F978-3-319-02961-0_46.pdfPublished version522.88 kB
Adobe PDF
Full metadata record
This paper proposes a new sclera vessel recognition technique. The vessel patterns of sclera are unique for each individual and this can be utilized to identify a person uniquely. In this research we have used a time adaptive active contour-based region growing technique for sclera segmentation. Prior to that, we have made some tonal and illumination correction to get a clearer sclera area without the distributing vessel structure. This is because the presence of complex vessel structures occasionally affects the region-growing process. The sclera vessels are not prominent in the images, so in order to make them clearly visible, a local image enhancement process using a Haar high pass filter is incorporated. To get the total orientation of the vessels, we have used Orientated Local Binary Pattern (OLBP). The OLBP images of each class are used for template matching for classification by calculating the minimum Hamming Distance. We have used the UBIRIS version 1 dataset for the experimentation of our research. The proposed approach has achieved high recognition accuracy employing the above-mentioned dataset. © Springer International Publishing 2013.
Please use this identifier to cite or link to this item: