This document describes an algorithm for detecting eyelids in images of human eyes. It begins with an introduction on the importance of eye shape detection. It then describes the steps of the algorithm in detail: detecting the iris center and radius, finding the upper eyelid contour by searching for luminance valley points along horizontal rows, estimating the lower eyelid from eye corners and iris shape, and fitting polynomial curves to the eyelid point sets. It also discusses experimental analysis, matching techniques, and references related work on eye detection.
Presentation by Pier Luca Lanzi at the Optimization by Building and Using Probabilistic Models (OBUPM 2008) workshop at the Genetic and Evolutionary Computation Conference (GECCO-2008)
Performance of genetic algorithm is flexible enough to make it applicable to a wide range of problems, such as the problem of placing N queens on N by N chessboard in order that no two queens can attack each other which is known as ‘n-Queens problem.
Lack of information about details of the problem made genetic algorithm confused in searching state space of the problem
SAL3D presentation - AQSENSE's 3D machine vision libraryAQSENSE S.L.
The 3D Shape Analysis Library (http://www.aqsense.com/products/sal3d) is the first hardware independent software architecture for range map and poing cloud processing, fully oriented to laser triangulation and 3D machine vision applications.
SAL3D means speed, accuracy, and reliability to machine builders, equipment manufacturers, system integrators, and volume end users demanding maximum flexibility and customization in their vision systems. Tools can be integrated as DLL's that allow developers access to third party components usable side by side with SAL's tools resulting in rapid development of highly complex processing tasks.
Presentation by Pier Luca Lanzi at the Optimization by Building and Using Probabilistic Models (OBUPM 2008) workshop at the Genetic and Evolutionary Computation Conference (GECCO-2008)
Performance of genetic algorithm is flexible enough to make it applicable to a wide range of problems, such as the problem of placing N queens on N by N chessboard in order that no two queens can attack each other which is known as ‘n-Queens problem.
Lack of information about details of the problem made genetic algorithm confused in searching state space of the problem
SAL3D presentation - AQSENSE's 3D machine vision libraryAQSENSE S.L.
The 3D Shape Analysis Library (http://www.aqsense.com/products/sal3d) is the first hardware independent software architecture for range map and poing cloud processing, fully oriented to laser triangulation and 3D machine vision applications.
SAL3D means speed, accuracy, and reliability to machine builders, equipment manufacturers, system integrators, and volume end users demanding maximum flexibility and customization in their vision systems. Tools can be integrated as DLL's that allow developers access to third party components usable side by side with SAL's tools resulting in rapid development of highly complex processing tasks.
Iris Localization - a Biometric Approach Referring Daugman's AlgorithmEditor IJCATR
In general, there are many methods of biometric identification. But the Iris
recognition is most accurate and secure means of biometric identification. Iris has
many properties which makes it ideal biometric identification. There are many
methods used to identify the Iris location. To locate Iris many traditional methods are
used. In this we proposed such methods which can identify Iris Center(IC) as well as
localize its center. In this paper we are proposing a method which can use novel IC
localization method on the fact that the elliptical shape (ES) of Iris varies according to
the rotation of eye movement. In this paper various IC locations are generated and
stored in database. Finally the location of IC is detected by matching the ES of the Iris
of input eye image withes candidates in DB. In this paper we are comparing different
methods for Iris localization.
OPTICAL MICROSCOPY AND COORDINATE MEASURING MACHINE sangeetkhule
Introduction
Working principle
Classification
Construction and working
Different types of an optical scope
Process capabilities and analysis
Testing
Process parameters
Components and machine structure
Confocal laser scanning microscopy
Microscopic
Advantages
Applications
Advancement in CMM
Machine characteristics
Process parameters of CMM
Animation video
Research papers
Bar graphs and tables
Conclusion
References
IRIS BIOMETRIC RECOGNITION SYSTEM EMPLOYING CANNY OPERATORcsitconf
Biometrics has become important in security applications. In comparison with many other
biometric features, iris recognition has very high recognition accuracy because it depends on
iris which is located in a place that still stable throughout human life and the probability to find
two identical iris's is close to zero. The identification system consists of several stages including
segmentation stage which is the most serious and critical one. The current segmentation
methods still have limitation in localizing the iris due to circular shape consideration of the
pupil. In this research, Daugman method is done to investigate the segmentation techniques.
Eyelid detection is another step that has been included in this study as a part of segmentation
stage to localize the iris accurately and remove unwanted area that might be included. The
obtained iris region is encoded using haar wavelets to construct the iris code, which contains
the most discriminating feature in the iris pattern. Hamming distance is used for comparison of
iris templates in the recognition stage. The dataset which is used for the study is UBIRIS
database. A comparative study of different edge detector operator is performed. It is observed
that canny operator is best suited to extract most of the edges to generate the iris code for
comparison. Recognition rate of 89% and rejection rate of 95% is achieved.
IRIS BIOMETRIC RECOGNITION SYSTEM EMPLOYING CANNY OPERATORcscpconf
Biometrics has become important in security applications. In comparison with many other biometric features, iris recognition has very high recognition accuracy because it depends on
iris which is located in a place that still stable throughout human life and the probability to find two identical iris's is close to zero. The identification system consists of several stages including
segmentation stage which is the most serious and critical one. The current segmentation methods still have limitation in localizing the iris due to circular shape consideration of the
pupil. In this research, Daugman method is done to investigate the segmentation techniques. Eyelid detection is another step that has been included in this study as a part of segmentation
stage to localize the iris accurately and remove unwanted area that might be included. The obtained iris region is encoded using haar wavelets to construct the iris code, which contains
the most discriminating feature in the iris pattern. Hamming distance is used for comparison of iris templates in the recognition stage. The dataset which is used for the study is UBIRIS database. A comparative study of different edge detector operator is performed. It is observed that canny operator is best suited to extract most of the edges to generate the iris code for comparison. Recognition rate of 89% and rejection rate of 95% is achieved.
Iris Localization - a Biometric Approach Referring Daugman's AlgorithmEditor IJCATR
In general, there are many methods of biometric identification. But the Iris
recognition is most accurate and secure means of biometric identification. Iris has
many properties which makes it ideal biometric identification. There are many
methods used to identify the Iris location. To locate Iris many traditional methods are
used. In this we proposed such methods which can identify Iris Center(IC) as well as
localize its center. In this paper we are proposing a method which can use novel IC
localization method on the fact that the elliptical shape (ES) of Iris varies according to
the rotation of eye movement. In this paper various IC locations are generated and
stored in database. Finally the location of IC is detected by matching the ES of the Iris
of input eye image withes candidates in DB. In this paper we are comparing different
methods for Iris localization.
OPTICAL MICROSCOPY AND COORDINATE MEASURING MACHINE sangeetkhule
Introduction
Working principle
Classification
Construction and working
Different types of an optical scope
Process capabilities and analysis
Testing
Process parameters
Components and machine structure
Confocal laser scanning microscopy
Microscopic
Advantages
Applications
Advancement in CMM
Machine characteristics
Process parameters of CMM
Animation video
Research papers
Bar graphs and tables
Conclusion
References
IRIS BIOMETRIC RECOGNITION SYSTEM EMPLOYING CANNY OPERATORcsitconf
Biometrics has become important in security applications. In comparison with many other
biometric features, iris recognition has very high recognition accuracy because it depends on
iris which is located in a place that still stable throughout human life and the probability to find
two identical iris's is close to zero. The identification system consists of several stages including
segmentation stage which is the most serious and critical one. The current segmentation
methods still have limitation in localizing the iris due to circular shape consideration of the
pupil. In this research, Daugman method is done to investigate the segmentation techniques.
Eyelid detection is another step that has been included in this study as a part of segmentation
stage to localize the iris accurately and remove unwanted area that might be included. The
obtained iris region is encoded using haar wavelets to construct the iris code, which contains
the most discriminating feature in the iris pattern. Hamming distance is used for comparison of
iris templates in the recognition stage. The dataset which is used for the study is UBIRIS
database. A comparative study of different edge detector operator is performed. It is observed
that canny operator is best suited to extract most of the edges to generate the iris code for
comparison. Recognition rate of 89% and rejection rate of 95% is achieved.
IRIS BIOMETRIC RECOGNITION SYSTEM EMPLOYING CANNY OPERATORcscpconf
Biometrics has become important in security applications. In comparison with many other biometric features, iris recognition has very high recognition accuracy because it depends on
iris which is located in a place that still stable throughout human life and the probability to find two identical iris's is close to zero. The identification system consists of several stages including
segmentation stage which is the most serious and critical one. The current segmentation methods still have limitation in localizing the iris due to circular shape consideration of the
pupil. In this research, Daugman method is done to investigate the segmentation techniques. Eyelid detection is another step that has been included in this study as a part of segmentation
stage to localize the iris accurately and remove unwanted area that might be included. The obtained iris region is encoded using haar wavelets to construct the iris code, which contains
the most discriminating feature in the iris pattern. Hamming distance is used for comparison of iris templates in the recognition stage. The dataset which is used for the study is UBIRIS database. A comparative study of different edge detector operator is performed. It is observed that canny operator is best suited to extract most of the edges to generate the iris code for comparison. Recognition rate of 89% and rejection rate of 95% is achieved.
3. Introduction
∗ The exact eye shape is a useful piece of input information
for applications like facial expression recognition, feature-
based face recognition and face modelling.
∗ The detection algorithm works in several steps:
∗ iris center and radius is estimated
∗ exact upper eyelid contour is detected by searching for
luminance valley points.
∗ lower eyelid is estimated from the eye corners coordinates
and iris.
4. Eye Contour Model
∗ The input data for the algorithm is a color image containing a single
human eye.
∗ The approximate scale and bounding box of the eye is considered
to be known.
∗ This normalization renders our detection techniques scale-
independent and permits us to work with absolute values of some
thresholds and parameters.
∗ Our eye contour model consists of upper lid curve (in cubic
polynomial), lower lid curve (in quadratic polynomial) and the iris
circle.
∗ The detection is performed in four steps:
• Iris center detection
• Iris radius estimation
• Upper eyelid contour detection
• Lower eyelid estimation
5. Cont…
∗ The iris center and radius detection is performed in
image’s red channel, which emphasizes the iris
border. This is due to the fact, that iris usually exhibits
low values of red (both for dark and light eyes), while
the surrounding pixels (sclera and skin) have
significantly higher red values.
6. IRIS DETECTION
∗ First, approximate iris center point is detected. The
central part of the eye image is checked for strong
highlight presence by comparing maximum pixel value
to a threshold.
∗ If strong highlight is found, the central area of eye
bounding box is scanned with a circular search window
with radius close to expected pupil (not iris) radius,
checking for several conditions:
∗ 1. The local minimum inside the search window should
not differ more than a given threshold from the global
minimum inside the eye bounding box (this makes
sure, that dark pixels are present in the search
window).
7. Cont…
∗ 2. The variance of pixel values inside the search
window should not be smaller than a certain portion
of global eye image variance (making sure, that both
dark and bright pixels are present inside the search
window);
∗ 3. The number of pixels darker than a given threshold
should be not less than a pre-defined value (checking,
that enough dark pixels are present inside the search
window);
8. Cont…
∗ All the locations, where these conditions are satisfied
are called the ”highlight locations”.
∗ Their coordinates are averaged to determine the
expected eye center (x0, y0).
∗ If no strong highlight is detected, a 5x5 minimum
filter is applied to eye area, to eliminate weak
highlight.
10. Accurate Iris Center Detection
∗ Ahlberg developed an algorithm in 1999 where two
assumptions are taken on the expected eye
appearance:
1.the iris is approximately circular and
2.it is dark against the background.
∗ The iris center and radius are found by searching for a
circle, which lies on the border between dark pixels of
iris and bright pixel of the eye white.
14. Eyelid Detection
∗ The upper lid detection is performed in three
stages.First, a set of points that belong to upper
eyelid is found. Then, this point set is examined to
remove outliers. Finally, a cubic polynomial curve is
fitted to the correct eyelid points.
∗ The lower lid is estimated by fitting a quadratic curve
to the eye corners and the lower point of the iris circle
- a reasonable approximation of the eye contour.
15. Cont…
∗ The well-known methods of eye shape estimation are
based on using high spatial luminance gradient locations
(so-called edges) as the attractors for the eyelid curve.
∗ One problem is very noisy edge map and therefore huge
amount of spurious edges, even for a clear eye image.
∗ The second is possible absence or discontinuity of the
significant edges.
∗ In some cases the brightness transition from sclera to eye
border and further to skin is too smooth to be identified as
an edge by a conventional edge detector.
16. Cont…
∗ Examining the luminance values change along a single
horizontal row of the eye image shows that significant
local minima correspond to eye boundary points.
∗ We deduce that looking for brightness valley points
instead of edge points is a more appropriate way for eye
shape estimation.
∗ We detect the luminance valley points, that most likely
correspond to the eye border in each row of the eye
image. These points are the points of significant local
minima of the horizontal luminance profiles.
18. Eye Opening Height
∗ The eye opening height is determined by scanning the
image iris area vertically from top to bottom calculating
each line’s average red value:
∗ Iy is the set of x-coordinates from the y line, that lie inside
the iris circle.
∗ |Iy| is the number of elements in the Iy set and
∗ R(x, y) is the red value of image pixel with (x, y)
coordinates.
∗ The area of low h(y) values indicate the area of visible (not
occluded by eyelids) iris area - the eye opening height.
19. Eye border points set construction
∗ After the eye opening height is known, the lines of
visible iris area are scanned outwards from the iris
borders in search for points that satisfy one
conditions:
∗ the point is a start of sharp luminance increase (we
have reached skin).
21. Erroneous and outlier points removal
∗ The border points set can contain outliers and erroneous
points, that would deviate fitted curve from the real eye
contour.
∗ To eliminate these outliers, two straight lines are fitted to the
left and right halves of the points set by the means of the
Hough transform .
∗ The Hough transform is known to be robust to imperfect data
and noise.
∗ Hough transform produces a set of lines, that pass through at
least 30% of the points in the subset.
∗ The line with maximum number of point lying closer than a
predefined distance ‘r’ is chosen as the principle line of the
subset.
∗ The points that lie too far are removed(outliers).
22. Eye border Points
(a) - initial eye border points set.
(b) - set with marked principle lines and
outliers removed.
(c) - a cubic polynomial curve
fitted to the final border points set.
23. Eyelid curves fitting
∗ Among the remaining points the leftmost and
rightmost are chosen to be the eye corners.
∗ All the points that lie above the line, connecting the
eye corners are treated as belonging to upper eyelid.
Finally, the upper iris border points are added to the
set and the lid curve is estimated by polynomial curve
fitting procedure.
∗ The lower lid is detected by fitting the eye corners
and the lower point of the iris circle with a quadratic
curve.
24. Experimental Analysis
∗ The algorithm was applied to images of
approximately 50 individuals taken under different
lighting conditions with different cameras and quality.
25. Matching
∗ Height of the opening
∗ Distance from right corner to left corner point.
∗ The hough transform distance ‘r’ .
26. Conclusion
∗ The method proposed is robust and sufficient
accuracy for face modelling application, while being
simple in implementation and fast in processing time
(especially compared with deformable models-based
methods).
27. References
1. Robust and Accurate Eye Contour Extraction, Vladimir
Vezhnevets & Anna Degtiareva, Graphics and Media
Laboratory, Moscow State University.
2. A New Method of Detecting Human Eyelids by
Deformable Templates ,Yuwen WU, Hong LIU, Hongbin
ZHA , National Lab. on Machine Perception, Peking
University, China.
3.A System for Face Localization and Facial Feature
Extraction, Jorgen Ahlberg,1999.