| Co-Investigator(Kenkyū-buntansha) |
KAMIYA Takeshi National Institution for Academic Degrees and University Evaluation, Faculty of University Evaluation and Research, Professor, 構・学位審査研究部, 教授 (70010791)
SHIMIZU Kayo Japan Women's Univ., Faculty of Science, Assistant, 理学部, 助手 (20318794)
|
|---|
| Budget Amount *help |
¥12,000,000 (Direct Cost: ¥12,000,000)
Fiscal Year 2003: ¥3,300,000 (Direct Cost: ¥3,300,000)
Fiscal Year 2002: ¥8,700,000 (Direct Cost: ¥8,700,000)
|
|---|
| Research Abstract |
With the progress of information technology, the computer-and the internet-systems have become indispensable in our daily lives. In addition to the growth of electronic commerce in economic life, cultural life begins to be changing by such as the introduction of distance education services, ore-learning. Under these beneficial circumstances, however, the need for security to protect the information from illegal accesses is proportionally increased. Any mistake in these systems could incur serious risks in our society. For this sake the technology of personal identification is increasingly recognized. Among other features, "face" is the most familiar element and less subject to psychological resistance as means of security control. Making use of this, since the 1970s in the field of the digital image processing technology, facial images have been one of the primary subjects to be examined. However there are still many technical subjects under study, such as methods for dealing with vari
… More
ation in the facial image and expression as a result of postural changes at acquisition as well as variations in lighting conditions.
The authors have proposed a new scheme using a multi-channel parallel JTC (RJTC) as a means of making better use of spatial parallelism through the use of a diffraction-type multi-level zone-plate array (MLAZPA) to extend a single channel JTC. A hybrid facial recognition system was then realized by adding pre and post digital processing. In a one-to-one correlation experiment conforming to the guideline of biometrics authentication, these two compact optical parallel correlators for facial recognition (COPaC II improved by COPaC: 20x24x43 cm^3, 6 kg, throughput time of 6.6 faces/s) achieved an error rate of less than 1% in terms of both the false rejection and false acceptance rates. From results, the COPaC II was verified to be applicable to computer login and home entrance security.
The vertical cavity surface emitting laser (VCSEL) proposed by K. Iga in 1977 has several remarks such as low-threshold operation, high electro-optical conversion efficiency, a circular emission distribution, easy mounting, and two dimensional integration, and as such is potentially applicable in optical communications and optical information processing. VCSELs are applicable in optical sensing using coherence with excellent single-mode characteristics. These tiny lasers can be expected to have an important device in many future technologies.
This research introduces a compact optical parallel facial image recognition system providing stable, high-speed, high-precision recognition constructed using this light source module. The design and trial fabrication of a two-dimensional light-source module is presented for application to an optical parallel correlator for facial recognition. The light-source module is composed of a vertical cavity surface emitting laser array and a multi-level zone-plate array as a collimating lens. This module is about 1/10 the size of conventional light source module and the optical parallel correlator has size of 16. 1x13x23 cm^3 and weight 4.4kg, and a throughput time of 19 faces/s. In an experimental evaluation of the system through one-to-one correlation using a database of 300 front facial images, the false match and false non-match rates were less than 1%. The optical system presented here is therefore robust toward a variety of changes in facial expressions and is highly applicability in security systems. These results will open the door to the next class of security systems. Less
|
