Content uploaded by Stéphane Derrode
Author content
All content in this area was uploaded by Stéphane Derrode on Jun 14, 2016
Content may be subject to copyright.
A low-cost video-based iris recognition system
Our system
1Replace the still camera by a video
camera to acquire a continuous
sequence of iris images.
2Select “on the fly” the workable
iris images for enrolment or
recognition.
Constraints
Low cost video camera
(webcam-type, 640 ×480).
Selection of iris images at the
frame rate (25 images/sec.)
⇒low complexity.
↬Embedded system
(PDA, mobile phone. . . )
First prototype
CMOS video camera
(ST Microelectronics)
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
pupil and iris localization results
US Patent Applications : L. Martin, G. Petitjean, S. Derrode, W. Ketchantang. Method and
device for locating a human iris in an eye image, No. 2008/0273,763, ST Microelectronics SA ;
Univ. Paul Czanne Aix-Marseille III. June 11, 2008.
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
Iris quality check
US Patent Applications : L. Martin, W. Ketchantang, S. Derrode, Method and device for selecting
images in a sequence of iris images received in a stream, No. 2008/0075,335, ST Microelectronics
SA ; Univ. Paul Czanne Aix-Marseille III. March, 27 2008.
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
Iris quality check
US Patent Applications : L. Martin, W. Ketchantang, S. Derrode, Method and device for selecting
images in a sequence of iris images received in a stream, No. 2008/0075,335, ST Microelectronics
SA ; Univ. Paul Czanne Aix-Marseille III. March, 27 2008.
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
Scale and rotation invariant coding of iris texture
Fourier-Mellin transform
Direct transform (k∈ℤet v∈ℝ)
ℳf𝜎(k,v) = 1
2𝜋∫∞
0∫2𝜋
0
f(r, 𝜃)r𝜎−iv e−ik 𝜃d𝜃
dr
r.
Inverse transform (pour r∈ℝ+
∗et 𝜃∈[0,2𝜋])
f(r, 𝜃) = ∫+∞
−∞ ∑
k∈ℤ
ℳf𝜎(k,v)r−𝜎+iv eik𝜃dv ,
Properties
Scale and rotation invariance can be achieved
Completeness of the description, which guaranties no loss of texture
information
Journal paper S. Derrode et F. Ghorbel, Robust and efficient Fourier-Mellin transform
approximations for invariant grey-level image description and reconstruction, Computer Vision and
Image Understanding, Vol. 83(1), pp. 57-78, juillet 2001.
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
Thanks and . . . projects
Project developed with
W. Ketchantang (SAGEM Security)
L. Martin (ST Microelectronics)
S. Bourennane (Institut Fresnel)
with a grant from (2003 →2008)
ST Microelectronics
PACA region, France
What next ?
Improve tracking (switching Kalman Filter)
Multi-images iris enrolment and recognition
Soft-based iris resolution improvement
↬To build a 2nd generation prototype
(embedded into a mobile)
Whom ?
N. Benletaief, PhD student
and support from . . .
Acknowledgment
YESS symposium comity,
French Embassy and MAE.
“Close-up” video sequences “On-the-fly” pupil/iris detection and localization “On-the-fly” iris quality assessment Iris signature extraction
Thanks and . . . projects
Project developed with
W. Ketchantang (SAGEM Security)
L. Martin (ST Microelectronics)
S. Bourennane (Institut Fresnel)
with a grant from (2003 →2008)
ST Microelectronics
PACA region, France
What next ?
Improve tracking (switching Kalman Filter)
Multi-images iris enrolment and recognition
Soft-based iris resolution improvement
↬To build a 2nd generation prototype
(embedded into a mobile)
Whom ?
N. Benletaief, PhD student
and support from . . .
Acknowledgment
YESS symposium comity,
French Embassy and MAE.