Publication list
2017 |
|||
| L. N. Darlow, B. Rosman. "Fingerprint Minutiae Extraction using Deep Learning," in the Proceedings of the World Congress on Internet Secruity, 2016, pp. 34-40. |  |
DOI | |
2016 |
|||
| L. N. Darlow, A. Singh, Y. Moolla, R. Ramokolo, R. van Wyk, N. Botha, L. Webb-Ray. "Damage invariant and high security acquisition of the internal fingerprint using optical coherence tomography," in the Proceedings of the World Congress on Internet Secruity, 2016, pp. 34-40. | |
Proceedings | |
| L. N. Darlow, J. Connan and A. Singh. "Performance analysis of a Hybrid fingerprint extracted from optical coherence tomography fingertip scans," in the Proceedings of the International Conference on Biometrics, 2016, pp. 1-8. | |
DOI | |
| Poster presented won Best Poster Award: |  |
||
| L. N. Darlow, L. Webb and N. Botha. "Automated spoof-detection for fingerprints using optical coherence tomography," Applied Optics, vol. 55, no. 13, 2016 pp. 3387-3396. | |
DOI | |
2015 |
|||
| L. N. Darlow and J. Connan. "Efficient internal and surface fingerprint extraction and blending using optical coherence tomography," Applied Optics, vol. 54, no. 31, 2015 pp. 9258-9268. | |
DOI | |
| L. N. Darlow and J. Connan. "A study on internal to surface fingerprint correlation using optical coherence tomography and internal fingerprint extraction," Journal of Electronic Imaging, vol. 24, no. 6, 2015. | |
DOI | |
| L. N. Darlow, J. Connan and S. S. Akhoury. "Internal fingerprint zone detection in optical coherence tomography fignertip scans," Journal of Electronic Imaging, vol. 24, no. 2, 2015. | |
DOI | |
| L. N. Darlow, S. S. Akhoury and J. Connan. "Internal fingerprint acquisition form optical coherence tomography fingertip scans," in the Proceedings of the Third International Conference on Digital Information, Networking, and Wireless Communications, 2015, 188-191. | |
DOI | |
| S. S. Akhoury and L. N. Darlow. "Extracting subsurface fingerprints using optical coherence tomography," in the Proceedings of the Third International Conference on Digital Information, Networking, and Wireless Communications, 2015, 184-187. | |
DOI | |
2014 |
|||
| L. N. Darlow, S. S. Akhoury and J. Connan. "A review of state-of-the-art speckle reduction techniques for optical coherence tomography fingertip scans," in the Proceedings of the Seventh International Conference on Machine Vision, SPIE, vol. 9445, no. 944523, 2014. | |
DOI | |
Current research: deep learning
My current research involves interrogating the functionality and learning dynamics of neural networks.
I wrote this blog post on GINN: Geometric Illustration of Neural Networks. It uses this web application to illustrate and track the decision boundaries of neural networks as they learn. The code is available here.
Optical coherence tomography and fingerprints
We exist in a world where security is of increasing concern. Fingerprints are an accepted solution to access control and personal identification. However, they are subject to:
- Damage: the fingertip is used every day and damage is inevitable
- Degradation: owing to its frequent use, fingertip skin degrades over time
- Distortion: during the touch based scanning process
- Spoofing: through dubious action to fake a fingerprint
To overcome this, we use an optical coherence tomography (OCT) scanner to obtain a touchless, higher-resolution, and 3D representation of the fingertip skin. My research entails obtaining a usable fingerprint from these.
What makes the OCT technology particularly interesting is that it provides unparalleled access to the internal fingerprint. This exists in a subsurface layer of skin and is a fingerprint that inherits the advantages of current surface fingerprints, yet is not subject to any of their known disadvantages.
My research entailed the following:
- A performance analysis of a hybrid fingerprint that is made up of the best quality constituent elements of the surface and internal fingerprints, both extracted from OCT scans. Download the paper:
"Performance Analysis of a Hybrid Fingerprint Extracted from Optical Coherence Tomography Fingertip Scans". - Automated spoof-detection of fingerprints. Download the paper:
"Automated spoof-detection for fingerprints using optical coherence tomography". - Internal and surface fingerprint extraction and blending for a hybrid fingerprint. Download the paper:
"Efficient internal and surface fingerprint extraction and blending using optical coherence tomography". - Detecting the location of the internal fingerprint in an OCT volume. Download the paper:
"Internal fingerprint zone detection in optical coherence tomography fingertip scans". - Internal fingerprint acquisition from 3D OCT scans. Download the papers:
"Internal fingerprint acquisition from optical coherence tomography fingertip scans".
"Extracting subsurface fingerprints using optical coherence tomography". - A study on the correlation between surface and internal fingerprints, using an internal fingerprint extraction algorithm I developed. Download the paper:
"A study on internal to surface fingerprint correlation using optical coherence tomography and internal fingerprint extraction" - Speckle noise reduction. I surveyed a number of speckle reduction techniques specific to OCT fingertip scans. Download the paper:
"A review of state-of-the-art speckle reduction techniques for optical coherence tomography fingertip scans".
Fingerprint alignment and blending
Albeit a work in progress, I am busy researching and developing algorithms for fingerprint alignment and blending. The paper entitled "Efficient internal and surface fingerprint extraction and blending using optical coherence tomography" introduces some of these algorithms in the context of OCT surface and internal fingerprints. I am porting these algorithms for use to strengthen existing fingerprint databases by combining several fingerprints into one super fingerprint.
Watch this space closely for publications and code. Feel free to contact me if you want to know more.
