Slide 1

Report
Video Based Palmprint Recognition
Chhaya Methani and Anoop M. Namboodiri
Camera Based Palmprint Recognition
Unconstrained Palm Imaging
• Challenges
• Towards civilian biometric applications
 Low resolution web cameras
 Unrestricted & unconstrained imaging
• Palmprint as a biometric
Unique
 Easy to capture
 User Convenience
High mobility
Easy integration simplifies use in a multibiometric setup
Video Based Palmprint Recognition
Figure showing difference in two images taken from the same user
 Background
 Contrast
 Noise
 Illumination
 Pose & scale
 Matching samples from the same user under different lighting
Detect and discard completely washed out samples
User 2
User 1
• Previous Work
 Hand extraction from background has been attempted [1]
 Pixel noise is dealt by selection of good features viz. Gabor[2]
 Noisy artefacts like blur, low image quality, shadow effects etc
 Illumination variations addressed by Intensity Normalization[2]
 Not effective for low contrast or fully saturated pixels
 Scale variations[1]
Hand extraction from background
 Pose variations[3]
 Accuracies can be improved by imbibing illumination invariance
Aplplications of Camera Based Palmprint Imaging a) Access Control ,b) Multibiometric Image Capture c)Mobile unlocking
[1] JDoublet, et al. “Contactless hand Recognition Using Shape and Texture Features”, ICSP 2006
[2] Zhang et al. “Online Palmprint Recognition”, PAMI 2003
[3] Methani,Namboodiri. “Pose Invariant Palmprint Recognition Recognition”, ICB 2009
Video Capture & Pre-Processing
Registration
FTA
Base
Image +
2
12.75%
0
12.75%
0
Paramet
er Sets
Base
Image
Score
Fusion
T0
FTA
Base
Image +
6
FTA
Base
image +
10
FTA
24.30
0
19.25%
0
36.19%
0
13.99%
0
4.70%
0
5.79%
0
T1
7.64%
6.57%
8.07%
8.09%
5.35%
7.58%
4.5%
8.09%
T2
14.36%
9.44%
5.82%
13.15%
4.64%
10.9%
3.62%
11.8%
T3
3.69%
11.8%
3.40%
16.86%
1.9%
13.99%
7.75%
14.5%
• Change in pose causes difference in reflection on different parts of the
hand
Can be used to advantage by using more frames with different lighting
The user is asked to move the palm a little while recording the video to
capture lighting variations
• Completely washed out samples are rejected by measuring texture
content of the image, viz mean & variance of edge response
Frame Combination & Matching
• Frame combination is preferred over super resolution due to high computational
complexity of the latter and a lack of robust landmark points in the low textured palm
 Registration usually done by finding corresponding landmark points in the two textures
The transformations then found is used to align the two views
images
 In the absence of robust keypoints on palm, the entire edge map is used to iteratively come closer to •the
Average rule was experimentally found to perform better
edge map of the consecutive frame
Since the motion of hand is smooth, any two frames can be aligned by adding transformations
of consecutive frames
Frame 2
Technique used
Base Image
Base Image + 2
Base Image + 6
Base image + 10
Max. rule
18.24%
16.05%
19.34%
26.86%
Second Max. rule
18.24%
15.03%
15.48%
19.39%
Average value
18.24%
10.94%
9.55%
9.50%
Frame 3
• Matching: matching is done using Gabor filter responses.
 Euclidean distance and gradient direction gives the nearest point in the consecutive frame
This is the approximate corresponding point
The same process is followed for each point on the edge map
Based on all the point matches, a transformation is computed and matches improved iteratively
 Matching the entire pattern ensures a good matching
Observations and Conclusions
Results and Experiments
• First row shows results of score fusion as more frames are added
• Rows 2-5 compare results while increasing quality control (threshold of expected minimum
texture) for each of the combinations formed by adding 0, 2, 6 and 10 images respectively
Palm line variation with change in view
• Important to overlay images separated in timeline before combining
Frame 1
• Preprocessing
 Detect unique frames based on camera capture rate using background substraction
 Correct in-plane rotation & extract palm image Correcting for in-plane rotations and palm extraction
 Compute edge map
 Reject samples having very low texture;
identified by the quality of the edge map obtained
a)Two images from same user showing variation with illiumination
b)completely saturated sample
• Video, instead of a single image, can capture more variations on the
user’s palm
 Only when image is parallel to imaging plane
• Video Capture
 Video of duration approx 2 sec is recorded
 Users are advised to hold the hand loosely and allow some motion of hand to
instigate illumination variation with view variation
 The background and the camera is fixed, with no restrictions on pose lighting etc
 Few frames from a recorded video
• Pose and illumination Invariance:
 Two fold challenge
Linear axis and Semilog axis ROC curve showing improvement in accuracies using multiple frames
•Dataset: 100 users, 6 videos each
 640x480 logitech web camera used
• Combining 11 frames in the feature domain (using
Gabor filter response as features) takes 1.4 secs on
MATLAB
•Parameters used for the Gabor filter experimentally
determined to be the follwing: Window size = 27, var =
6.4, freq = 0.08
• 3528 genuine comparisons and 1,75,065 imposter
comparisons are made
•Thresholds used on mean and variance of Gabor
response of images to reject “bad” samples before
running the algorithm
ROC curve on semilog
axis after rejecting
samples using threshold
T3 as
shown in the Table.
It can be seen that the
unusual drop in ROC is
lost, indicating that
certain washed out
samples in data caused
the problem
Center for Visual Information Technology
International Institute of Information Technology Hyderabad
http://cvit.iiit.ac.in
• Reason for average rule working better than max or second max seems to be the
inexact registration
• Slow rise in GAR(Genuine Accept Rate)initially in ROC curve indicates presence of some
imposter scores before genuine scores
 Indicates washed out samples
• These samples are then rid of when thresholding is applied, saving the recognizer
computational efficiency otherwise spent on “bad samples”
Conclusions and Future Work
• Video Based Palmprint Recognition results in better accuracies than when single image is used
due to the presence of extra information in the additional frames
• EER reduced from 12.79% to 4.70% on frame combination and further to 1.9% on removing the
washed out samples
• As a follow up to this work, an improvement in registration process shall improve combination
strategies and hence result in better strategies

similar documents