下載/瀏覽Download
Download
Report
Transcript 下載/瀏覽Download
KinectFaceDB : A Kinect Database
for Face Recognition
Rui Min, Neslihan Kose, and Jean-Luc Dugelay, Fellow, IEEE
IEEE TRANSACTIONS ON SYSTEMS, MAN, AND
CYBERNETICS: SYSTEMS, VOL. 44, NO. 11,
NOVEMBER 2014
Presenter : Siang Wang
Advisor :
Dr. Yen - Ting Chen
Date :
2015.03.25
2
Outline
Introduction
Review Of 3-D Face Databases
Kinect Face Database
Benchmark Evaluation
Data Quality Assessment Of KinectFaceDB And
FRGC
Conclusion
3
Introduction
The emerging RGB-D cameras such as the
Kinect sensor have been successfully applied
to many 3-D based applications.
•
•
•
•
Efficiency
Low-cost
Ease of RGB-D mapping
Multimodal sensing
4
Introduction
The adoption of this powerful new sensor for
face recognition has been mostly overlooked
due to the lack of a standard database for
testing.
5
Introduction
Face recognition,
the least intrusive biometric technique
applied to commercial and law enforcement
applications.
6
Introduction
That face recognition methods exploiting 3D cues are more efficient than 2-D-based
methods.
7
Introduction
Main contribution
A complete multimodal face database-based on
the Kinect is built and thoroughly described.
Extensive experiments are conducted for the
benchmark evaluations of 2-D, 2.5-D, and 3-D
based recognition using standard face
recognition methods.
8
Introduction
Main contribution
Recognition results on both the KinectFaceDB
and the FRGC
9
Review Of 3-D Face Databases
Face database is being proposed for two
main purposes
To test face recognition algorithms robust to one
or multiple facial variations
To assist with the development of face
recognition algorithms using a specific data
modality
10
Review Of 3-D Face Databases
Comparison between the 3-D & 2-D Face
Database
In comparison to the large number of 2-D face
databases, the number of available 3-D face
databases is relatively small.
3-D face recognition algorithms are less
affected by the illumination changes than 2-D
methods
11
Kinect Face Databases
The structure and the acquisition environment
of the proposed KinectFaceDB
Acquisition Process
The postprocessing steps
The potential usages of the proposed
KinectFaceDB
12
Kinect Face Databases
Database Structure
Fifty-two volunteers
13
Kinect Face Databases
Four types of data modalities are captured for
each identity
2-D RGB image
2.5-D depth map
3-D point cloud
RGB-D video sequence
14
Kinect Face Databases
nine facial variations
15
Kinect Face Databases
A protocol to record the RGB-D video
sequences
16
Kinect Face Databases
Acquisition Environment
Controlled indoor environment
Kinect (in front of participant)
White board (behind participant)
LED lamp(in front of participant)
17
Kinect Face Databases
Acquisition Process
Summarization of 3-D imaging procedure of the
Kinect
RGB and Depth Imaging From the Kinect
Converting to 3-D Face Data
RGB-D Alignment for Face Data
18
Kinect Face Databases
RGB and Depth Imaging From the Kinect
19
Kinect Face Databases
Converting to 3-D Face Data
20
Kinect Face Databases
The RGB and depth images are sampled
separately from two different cameras with a
displacement between them
21
Kinect Face Databases
RGB-D Alignment for Face Data
22
Kinect Face Databases
23
Kinect Face Databases
PostProcessing
Noise Removal
Facial Landmarking
24
Kinect Face Databases
Noise Removal
The point is too far
The point is too close
The point is in shadow cast by projecter
The surface reflects poor IR light
25
Kinect Face Databases
Facial Landmarking
Six anchor on the face
Left eye center
Right eye center
Nose-tip
Left mouth corner
Right mouth corner
Chin
26
Kinect Face Databases
Potential Database Usages in Addition to Face
Recognition
Facial Demographic Analysis
Three-Dimensional Face Modeling
27
Benchmark Evaluation
For 2-D and 2.5-D
PCA(Principal Component Analysis)
LBP(Local Binary Patterns)
SIFT(Scale-Invariant Feature Transform)
LGBP(Local Gabor Binary Pattern)
28
Benchmark Evaluation
29
Benchmark Evaluation
30
Benchmark Evaluation
For 3-D
ICP (iterative closest point)
TPS (thin-plate spline)
31
Benchmark Evaluation
32
Benchmark Evaluation
Fusion of RGB and Depth Face Data
combine both the RGB (2-D) and the depth
(2.5-D) face information from the Kinect in face
recognition
33
Benchmark Evaluation
34
Benchmark Evaluation
35
Data Quality Assessment Of KinectFaceDB
And FRGC
visually observe the 3-D
data quality differences
between the Kinect and
a high-quality laser
scanner
36
Data Quality Assessment Of KinectFaceDB
And FRGC
the depth accuracy and the depth resolution
Data quality
evaluate the identification/ verification
differences of 2.5-D/3-D faces captured by
the Kinect and a high-quality laser scanner
37
Data Quality Assessment Of KinectFaceDB
And FRGC
38
Data Quality Assessment Of KinectFaceDB
And FRGC
Kinect’s merits for face recognition
works in real time
3-D data provides complementary information
to the 2-D data, and can be easily integrated for
multimodal face recognition
with streaming 3-D video of the Kinect
39
Conclusion
complete multimodal (including well-aligned
2-D, 2.5-D, and 3-Dface data)
KinectFaceDB supplies a standard medium
to fill the gap between traditional face
recognition and the emerging Kinect
technology.
40
Conclusion
design of new algorithms and new facial
descriptors for the low-quality 3-D data
how to efficiently combine different data
modalities (RGB, depth,and 3-D) so as to
maximize the exploitation of the Kinect
41
Thanks for your attention