On the Feasibility of Interoperable Schemes in Hand Biometrics

Abstract

Personal recognition through hand-based biometrics has attracted the interest of many researchers in the last twenty years. A significant number of proposals based on different procedures and acquisition devices have been published in the literature. However, comparisons between devices and their interoperability have not been thoroughly studied. This paper tries to fill this gap by proposing procedures to improve the interoperability among different hand biometric schemes. The experiments were conducted on a database made up of 8,320 hand images acquired from six different hand biometric schemes, including a flat scanner, webcams at different wavelengths, high quality cameras, and contactless devices. Acquisitions on both sides of the hand were included. Our experiment includes four feature extraction methods which determine the best performance among the different scenarios for two of the most popular hand biometrics: hand shape and palm print. We propose smoothing techniques at the image and feature levels to reduce interdevice variability. Results suggest that comparative hand shape offers better performance in terms of interoperability than palm prints, but palm prints can be more effective when using similar sensors.

Feature exploration for biometric recognition using millimetre wave body images

“…Regarding shape-based recognition applied to other biometrics traits, we may find examples in hand [13,[25][26][27] and signature biometrics [28,29].…”

A novel hand reconstruction approach and its application to vulnerability assessment

“…In order to train the hand-shape generator and initialize the Uphill Simplex (i.e., compute the initialization parameters G, P , k,x), the GPDS2 DB [50] is used. This database comprises one sample of the right hand of 100 users, captured with a 60 dpi commercial scanner in one session.…”

“…The images used to train the hand generator and to compute the initialization parameters are taken from the GPDS2 database [50], while the real hand shape samples to be reconstructed are taken from the GPDS [20] and the UST [63] databases. It is important to notice that the images used to train the generator are independent and belong to completely different users than those being reconstructed.…”

Analysis of image processing, machine learning and information fusion techniques for contact-less hand biometrics

“…This way, the posterior processing of the images is eased and the intra-user variability derived from hand pose variations is reduced. Nevertheless, these capturing devices present some drawbacks such as shape distortion, discomfort or unnatural posture of the hands [55].Thus, these capturing devices evolved into scanners which requires the hand to be in contact with a flat surface but does not restrict the openess degree of the fingers nor the position [137,141,154,127,106,87]. Most of contact-based approaches guarantee uniform lighting conditions and well contrasted background with the aim of facilitating the segmentation process.…”

“…These methods include global thresholding, which is the easiest but most extended approach as commented above, and more sophisticated and well known methods such as Graph Cuts. Thresholding segmentation is a really simple solution which requires low computational resources but entail very specific capturing conditions to provide a good per- [153,189,137,156,65,70,141,27,51,111,195,72,11,194,155,154,115,7,56,5,53,100,58,197,8,55,196,128,3,179,147,127,106,87 Table 2. : Hand segmentation for biometric applications using visual spectrum images classified according to the nature of the testing images in terms of pose restrictions and environmental conditions.…”