ANKUSH KUMAR (M.Tech(CS))

Slides:



Advertisements
Similar presentations
Distinctive Image Features from Scale-Invariant Keypoints
Advertisements

Distinctive Image Features from Scale-Invariant Keypoints David Lowe.
Group Meeting Presented by Wyman 10/14/2006
Presented by Xinyu Chang
Computer vision: models, learning and inference Chapter 13 Image preprocessing and feature extraction.
CSE 473/573 Computer Vision and Image Processing (CVIP)
Instructor: Mircea Nicolescu Lecture 15 CS 485 / 685 Computer Vision.
Emerging biometrics Presenter : Shao-Chieh Lien Adviser : Wei-Yang Lin.
Face Recognition & Biometric Systems, 2005/2006 Face recognition process.
Biometrics & Security Tutorial 7. 1 (a) Please compare two different kinds of biometrics technologies: Retina and Iris. (P8:2-3)
SURF: Speeded-Up Robust Features
Patch Descriptors CSE P 576 Larry Zitnick
A Study of Approaches for Object Recognition
Iris Recognition Slides adapted from Natalia Schmid and John Daugman.
Feature extraction: Corners and blobs
Scale Invariant Feature Transform (SIFT)
Automatic Matching of Multi-View Images
Blob detection.
Iris Recognition By Mohammed, Ashfaq Ahmed. Introduction Iris Recognition is a Biometric Technology which deals with identification based on the human.
PhD Thesis. Biometrics Science studying measurements and statistics of biological data Most relevant application: id. recognition 2.
Face Recognition CPSC 601 Biometric Course.
Face Recognition Using Neural Networks Presented By: Hadis Mohseni Leila Taghavi Atefeh Mirsafian.
A survey of image-based biometric identification methods: Face, finger print, iris, and others Presented by: David Lin ECE738 Presentation of Project Survey.
Overview Introduction to local features
Interest Point Descriptors
Distinctive Image Features from Scale-Invariant Keypoints By David G. Lowe, University of British Columbia Presented by: Tim Havinga, Joël van Neerbos.
Computer vision.
05 - Feature Detection Overview Feature Detection –Intensity Extrema –Blob Detection –Corner Detection Feature Descriptors Feature Matching Conclusion.
Following the work of John Daugman University of Cambridge
Study and Implementation Of Iris Recognition Schemes By:Ritika Jain Under guidance of DR K R RAO UNIVERSITY OF TEXAS AT ARLINGTON.
Recognition and Matching based on local invariant features Cordelia Schmid INRIA, Grenoble David Lowe Univ. of British Columbia.
Overview Harris interest points Comparing interest points (SSD, ZNCC, SIFT) Scale & affine invariant interest points Evaluation and comparison of different.
Local invariant features Cordelia Schmid INRIA, Grenoble.
Reporter: Fei-Fei Chen. Wide-baseline matching Object recognition Texture recognition Scene classification Robot wandering Motion tracking.
March 10, Iris Recognition Instructor: Natalia Schmid BIOM 426: Biometrics Systems.
776 Computer Vision Jan-Michael Frahm Fall SIFT-detector Problem: want to detect features at different scales (sizes) and with different orientations!
Biometrics Stephen Schmidt Brian Miller Devin Reid.
Local invariant features Cordelia Schmid INRIA, Grenoble.
Distinctive Image Features from Scale-Invariant Keypoints Ronnie Bajwa Sameer Pawar * * Adapted from slides found online by Michael Kowalski, Lehigh University.
IRIS RECOGNITION. CONTENTS  1. INTRODUCTION  2. IRIS RECOGNITION  3. HISTORY AND DEVELOPMENT  4. SCIENCE BEHIND THE TECHNOLOGY  5. IMAGE ACQUISITION.
Iris Scanning By, rahul vijay 1. Introduction  Biometrics provides a secure method of authentication and identification.  Biometric identification utilises.
1 Iris Recognition Ying Sun AICIP Group Meeting November 3, 2006.
IRIS RECOGNITION SYSTEM
Distinctive Image Features from Scale-Invariant Keypoints David Lowe Presented by Tony X. Han March 11, 2008.
Jack Pinches INFO410 & INFO350 S INFORMATION SCIENCE Computer Vision I.
Overview Introduction to local features Harris interest points + SSD, ZNCC, SIFT Scale & affine invariant interest point detectors Evaluation and comparison.
Non-Ideal Iris Segmentation Using Graph Cuts
1 Machine Vision. 2 VISION the most powerful sense.
Presented by David Lee 3/20/2006
ATM WITH AN EYE. An automated teller machine (ATM) or automatic banking machine (ABM) is a computerised telecommunications device that provides the clients.
Recognizing specific objects Matching with SIFT Original suggestion Lowe, 1999,2004.
Blob detection.
IRIS RECOGNITION 1 CITY ENGINEERING COLLEGE Technical Seminar On “IRIS RECOGNITION” By NANDAN.T.MURTHY 1CE06EC043.
SIFT Scale-Invariant Feature Transform David Lowe
ATM WITH AN EYE Suresh kumar.
Presented by David Lee 3/20/2006
Lecture 07 13/12/2011 Shai Avidan הבהרה: החומר המחייב הוא החומר הנלמד בכיתה ולא זה המופיע / לא מופיע במצגת.
Distinctive Image Features from Scale-Invariant Keypoints
Scale Invariant Feature Transform (SIFT)
PRESENTED BY Yang Jiao Timo Ahonen, Matti Pietikainen
FACE RECOGNITION TECHNOLOGY
SURF: Speeded-Up Robust Features
Following the work of John Daugman University of Cambridge
CAP 5415 Computer Vision Fall 2012 Dr. Mubarak Shah Lecture-5
SIFT keypoint detection
SIFT.
ECE734 Project-Scale Invariant Feature Transform Algorithm
SIFT SIFT is an carefully designed procedure with empirically determined parameters for the invariant and distinctive features.
Recognition and Matching based on local invariant features
Presentation transcript:

ANKUSH KUMAR (M.Tech(CS)) Iris Recognition Using SURF. By ANKUSH KUMAR (M.Tech(CS)) 211CS2279 April 15, 2017 ANKUSH KUMAR

Contents Why Iris Recognition scores over others? Anatomy of the Human Eye The Iris Iris Recognition : An Overview The process Segmentation Normalization Feature Encoding & Matching Feature (Image Descriptor) SURF Matching Methods Applications References April 15, 2017 ANKUSH KUMAR

Biometric Features Can be used to uniquely identify individuals Face Fingerprint Handprint Voice Iris April 15, 2017 ANKUSH KUMAR

Why Iris Recognition scores over others? Has highly distinguishing texture Right eye differs from left eye Twins have different iris texture Not trivial to capture quality image + Works well with cooperative subjects + Used in many airports in the world April 15, 2017 ANKUSH KUMAR

Anatomy of the Human Eye Eye = Camera Cornea bends, refracts, and focuses light. Retina = Film for image projection (converts image into electrical signals). Optical nerve transmits signals to the brain. April 15, 2017 ANKUSH KUMAR

Iris Iris is the area of the eye where the pigmented or coloured circle, usually brown, blue, rings the dark pupil of the eye. Example of 10 Different People Iris April 15, 2017 ANKUSH KUMAR

Proposed Iris Recognition Systems John Daugman (1993) First and most well-known Wildes (1996) Boles (1998) Ma (2004) April 15, 2017 ANKUSH KUMAR

Analysis & Recognition Image Capture Iris Segmentation Feature Extraction Matching April 15, 2017 ANKUSH KUMAR

Typical iris system configuration Uniform distribution Stored templates Reject Pre processing Feature-extraction Identification Verification Accept Iris scan 2d image capture Iris localization Transform representation comparison enrolment Authentication April 15, 2017 ANKUSH KUMAR

Iris Recognition systems The iris-scan process begins with a photograph. A specialized camera, typically very close to the subject, not more than three feet, uses an infrared image to illuminate the eye and capture a very high-resolution photograph.This process takes 1 to 2 seconds. April 15, 2017 ANKUSH KUMAR

Preprocessing Image acquisition - Focus on high resolution and quality - Moderate illumination - Elimination of artifacts Image localization Adjustments for imaging contrast, illumination and camera gain April 15, 2017 ANKUSH KUMAR

Iris Recognition System Localization Acquisition IrisCode Gabor Filters Polar Representation Image Demarcated Zones April 15, 2017 ANKUSH KUMAR

Iris Segmentation Objective : To isolate the actual iris region in a digital eye image Can be approximated by two circles the iris/sclera boundary the iris/pupil boundary(interior to former) Depends on the image quality Ex :- persons with darkly pigmented irises will present very low contrast between the pupil and iris region if imaged under natural light April 15, 2017 ANKUSH KUMAR

Iris Localization Next, we must detect the outer boundary Use canny edge detector and Hough transform April 15, 2017 ANKUSH KUMAR

Normalization Once the iris region is successfully segmented from an eye image, the next stage is to transform the iris region so that it has fixed dimensions in order to allow comparisons The normalization process will produce iris regions, which have the same constant dimensions Two photographs of the same iris under different conditions will have characteristic features at the same spatial location April 15, 2017 ANKUSH KUMAR

The normalization process for two images of the same iris taken under varying conditions April 15, 2017 ANKUSH KUMAR

Daugman’s Rubbersheet Model Each pixel (x,y) is mapped into polar pair (r, ). Circular band is divided into 8 subbands of equal thickness for a given angle. Subbands are sampled uniformly in and in r. Sampling = averaging over a patch of pixels. April 15, 2017 ANKUSH KUMAR

Iris code generation April 15, 2017 ANKUSH KUMAR

Image Descriptor(feature) SIFT(Scale Invariant Feature Transform) GLOH (Gradient Location and Orientation Histogram) HOG (Histogram of oriented gradients) LESH (Local Energy based Shape Histogram) SURF (Speeded Up Robust Feature) Interest point detection Descriptor April 15, 2017 ANKUSH KUMAR

April 15, 2017 ANKUSH KUMAR

Detection Lxx(x,σ) is the Laplacian of Gaussian of the image Hessian-based interest point localization Lxx(x,σ) is the Laplacian of Gaussian of the image It is the convolution of the Gaussian second order derivative with the image Lindeberg showed Gaussian function is optimal for scale-space analysis April 15, 2017 ANKUSH KUMAR

Detection cont… Approximated second order derivatives with box filters (mean/average filter) April 15, 2017 ANKUSH KUMAR

Detection cont… Scale analysis with constant image size 9 x 9, 15 x 15, 21 x 21, 27 x 27  39 x 39, 51 x 51 … 1st octave 2nd octave April 15, 2017 ANKUSH KUMAR

Description Orientation Assignment Circular neighborhood of radius 6s around the interest point (s = the scale at which the point was detected) April 15, 2017 ANKUSH KUMAR

Description DESCRIPTOR COMPONENT April 15, 2017 ANKUSH KUMAR

Matching Fast indexing through the sign of the Laplacian for the underlying interest point The sign of trace of the Hessian matrix Trace = Lxx + Lyy Either 0 or 1 (Hard thresholding, may have boundary effect …) In the matching stage, compare features if they have the same type of contrast (sign) April 15, 2017 ANKUSH KUMAR

Analysis SURF is good at SURF is poor at handling serious blurring handling image rotation SURF is poor at handling viewpoint change handling illumination change SURF describes image faster than SIFT by 3 times SURF is not as well as SIFT on invariance to illumination change and viewpoint change April 15, 2017 ANKUSH KUMAR

Matching Methods Hamming Distance The Hamming distance gives a measure of how many bits are the same between two bit patterns. In comparing the bit patterns X and Y, the Hamming distance, HD, is defined as the sum of disagreeing bits (sum of the exclusive-OR between X and Y) over N, the total number of bits in the bit pattern. April 15, 2017 ANKUSH KUMAR

Matching Methods Normalized Correlation   April 15, 2017 ANKUSH KUMAR

An illustration of the feature encoding process. ANKUSH KUMAR

An illustration of the shifting process. The lowest Hamming distance, in this case zero, is then used since this corresponds to the best match between the two templates. April 15, 2017 ANKUSH KUMAR

April 15, 2017 ANKUSH KUMAR Successful images

Observations Two IrisCodes from the same eye form genuine pair => genuine Hamming distance. Two IrisCodes from two different eyes form imposter pair => imposter Hamming distance. Bits in IrisCodes are correlated (both for genuine pair and for imposter pair). The correlation between IrisCodes from the same eye is stronger. April 15, 2017 ANKUSH KUMAR

Pros Iris is currently claimed and perhaps widely believed to be the most accurate biometric, especially when it comes to FA rates. Iris has very few False Accepts (the important security aspect). It maintains stability of characteristic over a lifetime. Iris has received little negative press and may therefore be more readily accepted. The fact that there is no criminal association helps. The dominant commercial vendors claim that iris does not involve high training costs. April 15, 2017 ANKUSH KUMAR

Cons There are few legacy databases. Though iris may be a good biometric for identification, large-scale deployment is impeded by lack of installed base. Since the iris is small, sampling the iris pattern requires much user cooperation or complex, expensive input devices. The performance of iris authentication may be impaired by glasses, sunglasses, and contact lenses; subjects may have to remove them. The iris biometric, in general, is not left as evidence on the scene of crime; no trace left. April 15, 2017 ANKUSH KUMAR

Conclusion The iris is an ideal biometric feature for human identification Although relatively young, the field of iris recognition has seen some great successes Commercial implementations could become much more common in the future April 15, 2017 ANKUSH KUMAR

References “How iris recognition works” by J. Daugman, Proceedings of 2002 International Conference on Image Processing, Vol. 1, 2002. “Recognition of Human Iris Patterns for Biometric Identification” by Libor Masek, The University of Western Australia, 2003 http://en.wikipedia.org/wiki/SURF Patch Descriptors, by :- Larry Zitnick (larryz@microsoft.com) “SURF: Speeded Up Robust Features”. IEEE Explore By Herbert Bay. April 15, 2017 ANKUSH KUMAR

THANKS April 15, 2017 ANKUSH KUMAR