Presentation is loading. Please wait.

Presentation is loading. Please wait.

ECE 5367 – Presentation Prepared by: Adnan Khan Pulin Patel Carlos Reyes Abbas Zaidi Pritesh Patel.

Similar presentations


Presentation on theme: "ECE 5367 – Presentation Prepared by: Adnan Khan Pulin Patel Carlos Reyes Abbas Zaidi Pritesh Patel."— Presentation transcript:

1 ECE 5367 – Presentation Prepared by: Adnan Khan Pulin Patel Carlos Reyes Abbas Zaidi Pritesh Patel

2

3 What is … Automatic recognition of a person using distinguishing traits Measurable: quantifying traits Robustness: extent to which trait changes over time Distinctiveness: measure of variations among patterns over a population

4 Biometrics as Authentication… Something you have – card, token Something you know – PIN, password Increasing level of security Increasing convenience Something you are – Biometric

5 Definition (contd.) Identification Identification Who is Who? “one-to-many” search (1:N) Used in identifying criminals Verification Verification Is this Him/Her? “one-to-one” search (1:1) Used for physical or computer access

6 Biometric Technologies… Fingerprint Fingerprint Fingerprint pattern read & matched electronically Iris Recognition Iris Recognition Measurement of the Iris pattern of the eye Facial Recognition Facial Recognition Recording spatial geometry of distinguishing features Speaker/Voice Recognition Speaker/Voice Recognition Voice or speaker recognition uses vocal characteristics for identification

7 Facial Recognition Based on ability to recognize faces Nodal points 80 nodal points on the face E.g.: Distance between eyes, width of nose, Cheekbones, Chin etc “Face-Print” – code of the face Numerical Code String of numbers

8 Facial Recognition (contd.) Algorithm 1) Capture Image 2) Find Face on the image 3) Extract Features (to generate template) 4) Compare Templates 5) Declare Matches Show Example

9 Fingerprint Formation of a unique signature Combination of Environmental and Genetic Factors : Position of foetus, density of amniotic fluid, DNA’s impact on each ridge The lowest mis-identification rate after Iris recognition Three main methods of capturing finger print : 1) Thermal Imaging 2) Capacitance Scanner 3) Optical Reader

10 Thermal Imaging Sensor The sensor uses the heat patterns of the finger to create slices of a person's fingerprint Each slice is then lined up using centering techniques

11 Capacitance and Optical Capacitance Sensing : The plates act as capacitors Finger acts as third plate Capacitance under ridge higher Capacitance under valley lower Optical Sensing : Uses Charged Coupled Devices (CCD). An Array of photosites which allow a current to flow through when hit by photons. Darker areas are valleys and lighter areas are ridges.

12 Analysis of the Prints Typica : 1) Minutiae Placement of ridges and valleys 2) Bifurcations Splitting of lines into two and their recombining

13 Iris Recognition Background Background Identifying persons without physical contact Used by the U.S. government, correctional facilities, nuclear utilities, banks, etc. The iris is the externally visible, colorful, donut-shaped organ surrounding the pupil of the eye.

14 Iris Recognition cont’d…(2) Uniqueness of the Iris Uniqueness of the Iris The iris is unique because of the morphogenesis of that organ. Stability of the Iris Stability of the Iris An iris is not normally contaminated with foreign material (eyeglasses or contact lenses.) It is not subject to deleterious effects of aging. The features of the iris remain stable and fixed from about one year of age until death.

15 Iris Recognition cont’d…(3) Hamming Distance Calculation Hamming Distance Calculation

16 Recognition Speed Recognition Speed Determined by the speed of the processor and size of the database (normally less than two seconds). Sunglasses and colored contact lenses pose no impediment to recognition. Biggest Advantages Biggest Advantages Accuracy with an order of magnitude greater than other biometric systems. Iris Recognition cont’d…(4)

17 Voice Recognition Automated process of identifying a specific individual's voice Sound signal is digitized Result is a simple yes/no decision Deployed in such areas as banking, account access, home PC and network access

18 Voice Recognition (cont…) Advantages Advantages Easy to use and require no special training No need to periodically manage/change passwords Relatively inexpensive compared to other biometrics

19 Questions?


Download ppt "ECE 5367 – Presentation Prepared by: Adnan Khan Pulin Patel Carlos Reyes Abbas Zaidi Pritesh Patel."

Similar presentations


Ads by Google