Rolled Equivalent, 3-D Fingerprint and Palm Print Scanner

The goal of this project is to advance the state of ten print rolled equivalent fingerprint capture, addressing the limitations of the current generation of technology. We will develop a scanner that is capable of capturing ten rolled equivalent fingerprints and two palm prints quickly, accurately and independently of an operator and is designed to function well even in high volume environments such as airports. The prints will be captured using a non-contact, 3-D technique, significantly enhancing the image quality of data collected over current technologies, and ultimately producing more efficient searches and better matching rates. The 3-D prints are "flattened" to produce 2-D rolled equivalent prints for identity checks in legacy Automated Fingerprint Identification System (AFIS) databases making the device backwards compatible. The proposed fingerprint capture system is based upon the principles of Structured Light Illumination (SLI) and uses the triangulation of light patterns between a digital projector and a digital camera. Structured light is a commonly used technique for making 3-D surface measurements for industrial inspection, and is able to make depth measurements on the order of 50 micrometers standard deviation. The technique is perfectly suited for fingerprintlpalm print scanning for a number of reasons. First, it is non-contact so minutiae points are easier to detect because there is no distortion in the print. Second, it is very scalable - able to work with an area the size of a fingerprint or an entire hand. Lastly, it captures the details of the fingerprint - the ridges, valleys, and minutiae in 3-D. It is precisely this 3-D structure that gives a fingerprint its unique latent print characteristics and allows for the formation of minutiae. Capturing this data in 3-D allows for a more accurate capture of the true object rather than a 2-D representation of a 3- D object and ultimately yields more accurate data. Unlike competing camera-based systems which acquire only 2-D photographs of what a fingerprint looks like, our system makes quantitative 3-D surface measurements that reveal actual ridge shape and height. The result is an extremely high quality image that when converted to 2-D, yields better image quality and minutiae detection results than 2-D scanners. It is our belief that because we can capture 3-D detail and measure the ridge height, that we may identi~' features not currently in use for differentiating between prints, and hence improve recognition performance. We will test this theory by developing a 3-D matching algorithm which will compare 3-D matching rates against 2-D matching rates. The University of Kentucky will partner with FlashScan3D and the University of Louisville on this project. The parties are currently working together on the Fast Capture Fingerprint/Palm Print project for the Department of Homeland Security (DHS). The goal of the Fast Capture project is to create a non-contact, structured light scanner that will capture all 10 fingerprints (nail-to-nail) and two palm prints in under 1 minute. University of Kentucky and FlashScan3D are also partnering on a Biometric Detector project for DHS, the objective of which is to develop a non-contact, single finger scanner. We feel that this solicitation represents a unique opportunity for the Department of Homeland Security to leverage not only the investment in the technical base of the Fast Capture project but the experience of the team and the technical breakthroughs we have been able to achieve. Because we are leveraging a technical base that already exists, a refined and optimized prototype can be completed in twelve months with a budget of $988,511 without a high degree of technical risk. This will allow us to focus a substantial part of our efforts on the development of a 3-D matching algorithm and evaluation of the 3-D data and scanner performance.