下載/瀏覽

Download Report

Transcript 下載/瀏覽

Distinctive Image Features from
Scale-Invariant Keypoints
David G. Lowe
International Journal of Computer Vision ,Volume 60 Issue 2,
Pages 91 – 110, 2004.
Presenter :JIA-HONG,DONG
Advisor : Yen- Ting, Chen
1
Outline





Introduction
Methodology
Recognition examples
Simulation and testing
Conclusion
2
Introduction
 Scale Invariant Feature Transform(SIFT)







Object and scene recognition
Video Tracking
Robotic mapping and navigation
Image stitching
3D modeling
Gesture recognition
Match moving
3
Introduction
 Advantage
 The features are invariant:





Image scaling
Image rotation
Illumination
Noise
Camera viewpoint
 High performance:
 High accuracy
 Near real-time
4
Introduction
 Major stages of computation:





Scale-space extrema detection
Keypoint localization
Orientation assignment
Keypoint descriptor
Matching features(nearest neighbor)
5
Methodology
 Detection of scale-space extrema
.....(1)
 Convolution
.....(2)
 image subtraction
.....(3)
6
Methodology
7
Methodology
.....(4)
.....(5)
.....(6)
(К-1) is a constant
factor σ2 is scale invariant as studied by Lindeberg (1994)
σ2Δ2G is maxima and minima image features as studied by Mikolajczyk(2002)
8
Methodology
9
Methodology
 Database
 32 real images:




Outdoor scenes
Human faces
Aerial photographs
Industrial images
10
Methodology
 Database
 Transformations:





Rotation
Scaling
Brightness
Contrast
Noise
11
Methodology
 Local extrema detection
12
Methodology
 Local extrema detection
13
Methodology
 Frequency of sampling in scale
14
Methodology
 Accurate keypoint localization
.....(3)
 Using Taylor expansion
.....(8)
 x=(x, y , σ )T is the offset from this point
.....(9)
15
Methodology
.....(10)
16
Methodology
17
Methodology
18
Methodology
 Orientation assignment
...(11)
.....(12)
L(x,y) is the sample image.
Θ(x,y) is orientation
m(x,y) is the gradient magnitude
19
Methodology
 Orientation histogram
 A region around the keypoint
 36 bins covering the 360 degree range of
orientations
 Added weighted
 A Gaussian window σ=1.5
20
Methodology
 The local image descriptor
21
Methodology
 Descriptor testing
22
Methodology
 Descriptor representation
 4x4 array of histograms
 8 orientation bins
 4x4x8 = 128 element feature vector
23
Methodology
 Keypoint matching
 Minimum Euclidean distance
24
Methodology
 Keypoint matching
25
Methodology
 Application to object recognition
 Need 3 features at least
 Higher probability
26
Recognition examples
 Descriptor testing
27
Recognition examples
 Sensitivity to affine change
28
Recognition examples
 Matching to large databases
29
Recognition examples
30
Recognition examples
31
Simulation and testing
32
Simulation and testing
33
Conclusion
 SIFT keypoints described are
particularly useful.
 A high-dimensional vector represents
the image gradients within a local
region of the image.
 Near real-time performance on
standard PC hardware.
34
Conclusion
 Systematic testing is needed on data
sets with full 3D viewpoint.
 Feature sets are likely to contain both
prior and learned features.
35
Thank you for your attention
36