Full metadata record
| DC Field | Value | Language |
|---|---|---|
| dc.contributor | Department of Computing | en_US |
| dc.contributor.advisor | Zhang, David (COMP) | - |
| dc.creator | Zhang, Kunai | - |
| dc.identifier.uri | https://theses.lib.polyu.edu.hk/handle/200/9505 | - |
| dc.language | English | en_US |
| dc.publisher | Hong Kong Polytechnic University | - |
| dc.rights | All rights reserved | en_US |
| dc.title | Online touch-based and touchless palmprint recognition | en_US |
| dcterms.abstract | Biometrics has been a popular method to verify personal identities. Face, fngerprint, iris, and palmprint are the most successful biometric technologies and have been used in door access control, online payment and etc. Among these technologies, palmprint recognition is relatively new but has excellent performance in terms of accuracy and anti-spoof ability. Palmprint recognition systems can be categorized into two types: touch-based systems and touchless systems.Traditional palmprint systems are touch-based, making use of the pegs and the cover of the device to make the hand pose and illumination under control. In this way, stable palmprint ROIs (region of interest) and stable features can be extracted accurately. As far as we know, touch-based palmprint systems can achieve a verifcation rate over 99.97% that outperforms any other biometrics. However, different palmprint devices of the same model have slightly different performances due to the inconsistency of the hardwares such as camera position, camera lens, light source and etc. The inconsistency problem is one of the reasons of the increase of false rejection rate (FRR). Therefore, it is important to detect and solve this problem. One of the solutions is to conduct image quality assessment. In addition to the inconsistency problem, there are raising concerns about the touch-based palmprint systems, such as the hygiene problem on the contact interface and leaving latent palmprints. Also, the touch-based nature makes the touch-based systems not convenient to use because users have to put their palms on a specifc interface with the help of the pegs. Compared with touch-based systems, touchless palmprint systems are more user-friendly but more challenging because the acquisition algorithm and feature extraction algorithm both have to be robust to the variation of illumination, background and hand deformation and poses. There are many ways to implement touchless palmprint recognition technology. One of the most popular ways is to use a smartphone because it has all the hardwares required for touchless palmprint recognition. The whole process can be done within a smartphone including image acquisition, image preprocessing, feature extraction and recognition result visualization. The smartphone can also invoke other applications according to the recognition result, like unlock the phone screen, authenticate an online payment and etc. In chapter 2 of this thesis, a calibration method is presented to solve the consistency problem of the touch-based palmprint systems. After calibration of the camera view, image defnition and image brightness, the recognition rate can be improved signifcantly. Then in chapter 3 and 4, methods of image sharpness adjustment are proposed to improve the recognition performance. An optimal range is computed by analyzing data ranging from clear images and defocused images. Results validate that if images are tuned to the optimal range, using Gaussian fltering and Gaussian inverse filtering, the performance of palmprint recognition can be improved. In chapter 5, to make touchless palmprint recognition possible on smartphones, a new acquisition method is proposed which uses the smartphone camera to locate the palmprint ROI accurately. Based on this method, a touchless palmprint dataset containing 2000 images from 200 palms is collected by a smartphone. A new feature extraction algorithm is also proposed making use of both the palmprint and inner knuckle textures. The proposed method outperforms other touchless algorithms and the testing results indicate that it can be applied for real-time palmprint recognition on smartphones. | en_US |
| dcterms.extent | xv, 65 pages : color illustrations | en_US |
| dcterms.isPartOf | PolyU Electronic Theses | en_US |
| dcterms.issued | 2018 | en_US |
| dcterms.educationalLevel | Ph.D. | en_US |
| dcterms.educationalLevel | All Doctorate | en_US |
| dcterms.LCSH | Hong Kong Polytechnic University -- Dissertations | en_US |
| dcterms.LCSH | Pattern recognition systems | en_US |
| dcterms.LCSH | Identification -- Automation | en_US |
| dcterms.LCSH | Biometric identification | en_US |
| dcterms.accessRights | open access | en_US |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| 991022141357803411.pdf | For All Users | 9.98 MB | Adobe PDF | View/Open |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/9505