Escolar Documentos
Profissional Documentos
Cultura Documentos
To
Next Generation Of security
Ascent 2K5 1
Image Processing
in
Bio-medical Field
Introduces
Presented By:
Miss. Jain Chetna A. Mr. Kothari Abhijeet D.
Guided By :
Dr.Mrs.S.D.Apte
Ascent 2K5 2
Image Processing
• Analysis of Graphical Images.
• Three Steps of Image Processing:
Input :- converts colors/shades into binary values.
Processing :- Image Enhancement, Normalization,
editing, filtering & data compression.
Output: display/printing processed image.
Ascent 2K5 3
IRIS Recognition
• Protected Internal Organ, Stable Throughout the Life.
• Iris Causes the muscles compression or expansion, as
pupil responds to changing lighting condition.
• Details of each Iris Phenotypically Unique
• Living Passport/ Living Password.
• Real Time, Recognition decisions made with High
Confidence even for Large Databases.
• Developed at Cambridge University by Dr. John
Daugman.
Ascent 2K5 4
How It Works
• User stands 1 meter from system.
• Contains 3 video cameras.
• 2 wide angle cameras for torso
• 3rd focuses on eye.
• Cameras uses visible & Infrared
(700-900nm) light.
• Circular grid is used as a guide to
encode the pattern.
Ascent 2K5 5
• Grid is overlaid on eyes image.
• Looks at pattern of light/dark
iris area
• 512 byte human bar code for
that person is created.
• This file size is small.
• System compares the bar code
with the version stored in a
computer database.
• Complete process takes about
2 seconds.
• Hamming Distance:
II(code A X-OR code B)∩mask A∩mask BII
II mask A∩ mask BII
Ascent 2K5 6
Eyes & Their Bar Patterns:
Here Black Color Represents ‘Logic 1 ‘
whereas White Color Represents ‘Logic 0’.
Ascent 2K5 7
Comparison
Ascent 2K5 8
Hough Transforms
Ascent 2K5 9
Advantages
Highly protected internal organ of the eye.
100% Accuracy and Reliability
Uniqueness: set by combinatorial complexity.
No effect of using contact lenses.
Patterns apparently stable throughout life.
Encoding and decision-making are tractable
Image analysis and encoding time: 1 second
Decidability Index = 7.3 to 11.4 seconds.
Search speed: 100,000 Iris Codes per second on
300MHz CPU.
Ascent 2K5 10
Disadvantages
Small target (1 cm) to acquire from a distance (1 m).
Obscured by eyelashes, lenses, reflections.
Partially occluded by eyelids.
Deforms non-elastically as pupil changes size.
Fake detection by “Eye Transplantation”.
Ascent 2K5 11
Famous Example
Ascent 2K5 13
Existing Applications
Password Authenticity Over Computer Networks
Ascent 2K5 14
Banking System
Ascent 2K5 15
Conclusion
Ascent 2K5 16
References
• Image Processing By: A. K. Jain.
• www.whatis.com
• www.google.com
• www.cl.cam.ac.uk
• www.accessexcellence.org
• www.melia.com
• www.iridiantech.com
• www.eyeticket.com
• www.nei.nih.gov
• www.webvision.med.utah.edu
• IEEE Report :Detecting Faces in Images: By Ming-
Hsuan Yang, David J. Kriegman & Narendra Ahuja.
Ascent 2K5 17
Ascent 2K5 18