(1 slide per page)
Transcript of (1 slide per page)
![Page 1: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/1.jpg)
Image Processing - Lesson 12
Segmentation
• Threshold Segmentation
• Local thresholding
• Edge thresholding
• Threshold using averaging
• Gradient Detectors
• Region Growing
• Split & Merge
• Shape Matching
• Shape Representation
![Page 2: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/2.jpg)
Segmentation
Image Segmentation = divide image into (continuous)regions or sets of pixels.
1) Region Based
2) Boundary Based
3) Edge Based
![Page 3: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/3.jpg)
Thresholding
Global Thresholding = Choose threshold T that separates object from background.
BackgroundObject
TGraylevel
# of
Pix
els
Image Histogram
![Page 4: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/4.jpg)
0 100 2000
500
1000
1500
Segmentation using Thresholding
Original Histogram
50 75
Threshold = 50 Threshold = 75
![Page 5: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/5.jpg)
0 100 2000
500
1000
1500
Original Histogram
21
Threshold = 21
![Page 6: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/6.jpg)
Thresholding a Grayscale Image
Original Image Thresholded Image
Threshold too low Threshold too high
![Page 7: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/7.jpg)
FMRI - Example
Original Image Threshold = 80
Threshold = 71 Threshold = 88
![Page 8: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/8.jpg)
Simple thresholding is not always possible:
Graylevel
# of
Pix
els
Image Histogram
1) Many objects at different gray levels.
2) Variations in background gray level.
3) Noise in image.
![Page 9: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/9.jpg)
Thresholding Example
Original
0 100 2000
50
100
150Histogram
Single Global Threshold
0 100 2000
50
100
150
T = 128
![Page 10: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/10.jpg)
Local Thresholding - 4 ThresholdsDivide image in to regions. Perform thresholding
independently in each region.
0 100 2000
50
100
0 100 2000
50
100
0 100 2000
50
100
0 100 2000
50
100
T = 80 T = 128 T = 188 T = 226
![Page 11: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/11.jpg)
Adaptive Thresholding
Every pixel in image is thresholded according to thehistogram of the pixel neighborhood.
0 50 1000
100
200
x-coordinate
Thre
shol
d Le
vel
T =
![Page 12: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/12.jpg)
Adaptive Thresholding - Example
Original
Global Threshold Adaptive Threshold
![Page 13: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/13.jpg)
Threshold Segmentation of Noisy Images
Noise inhibits localization of threshold.
Graylevel
# of
Pix
els
Image Histogram
Smooth image and obtain a histogram for whichthreshold is easily determined.
Graylevel
# of
Pix
els
Image Histogram
T
![Page 14: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/14.jpg)
Note: Smooth the image, not the histogram...
Graylevel
# of
Pix
els
Image Histogram
Graylevel
# of
Pix
els
Image Histogram
![Page 15: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/15.jpg)
Threshold using Average
Gray level
Gray level Histograms
no p
ixel
s
0 50 100 150 200 2500
5000
10000
![Page 16: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/16.jpg)
Threshold using Average
Gray level
no p
ixel
s
0 50 100 150 200 2500
50
100
0 50 100 150 200 2500
200
400
Gray level
Gray level Histograms
![Page 17: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/17.jpg)
Edge Based Segmentation
Edge ImageOriginal
ObjectBackground
![Page 18: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/18.jpg)
Edge Based Thresholding
Original Edge Pixels Edge Neighbors
0 50 100 150 200 2500
50
100
150
200
Edge Neighbors Histogram
![Page 19: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/19.jpg)
Thresholding Based on Boundary Characteristics
Original Threshold (T=182)l
0 100 2000
1000
2000
3000
Global Histogram
![Page 20: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/20.jpg)
Thresholding Based on Boundary Characteristics
Original Threshold (T=143)l
0 100 2000
5
10
15
20
Edge Neighborhood Histogram
![Page 21: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/21.jpg)
Region Growing
Define:
S = the set of pixels inside the region.
Q = queue of pixels to be checked.
(x0,y0) = a pixel inside the region.
Initialize: S = ∅
Q = { (x0,y0) }
1) Extract pixel P from queue Q
2) Add P to S.
3) For each neighbor P' of P:
if P' is "similar" to P and P' ∉ S then
add P' to Q.
4) If Q = ∅ then end, else return to 1.
Algorithm:
S = the extracted pixels of the region.Define what "similar" means.Problematic in small gradient regions.
![Page 22: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/22.jpg)
Region Growing - Example
Seed
![Page 23: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/23.jpg)
Region Growing - Examples
Color Segmentation
Texture Segmentation
Color + Texture Segmentation
![Page 24: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/24.jpg)
Watershed Threshold Algorithm
An Image can be viewed as a topographic map
T0
Tn
Thre
shol
d Le
vel
Image location
![Page 25: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/25.jpg)
Watershed Threshold Algorithm
Original T = 149 T = 150
Thre
shol
d Le
vel
T0
Tn
Image location
Initialize threshold at T0 that separates objects well.Determine connected components. Raise threshold, and detect pixels that pass threshold and belong to more than one connected component.Do not let objects merge. Set these pixels as object boundaries.
![Page 26: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/26.jpg)
Watershed Threshold Algorithm
![Page 27: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/27.jpg)
Watershed Threshold Algorithm
Original Watershed Boundaries
Watershed Markers Watershed Boundaries
Watershed Markers may be chosen manually or local global maximas (as above)
![Page 28: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/28.jpg)
Split & Merge Segmentation
2 Stage Algorithm:
Stage 1: SplitSplit image into regions using a Quad Tree representation.
Stage 2: MergeMerge "leaves" of the Quad Tree which are neighboring and "similar".
Original Split Split + Merge
![Page 29: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/29.jpg)
Quad Tree - RepresentationImage Quad Tree
1 2 3 4
1 2
3 4
Demo
![Page 30: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/30.jpg)
Quad Tree Representation
Original
Thresh = 0.20
Thresh = 0.40
Thresh = 0.55
![Page 31: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/31.jpg)
Split & Merge Example
Stage 1: Split
Stage 2: Merge
![Page 32: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/32.jpg)
Split & Merge Example
![Page 33: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/33.jpg)
Graph-Cut Segmentation
Wij
ji Wij
G = { V , E }
V = verticesE = Edges
V = image pixelsE = pixel similarity
Segmentation = Graph Partitioning
![Page 34: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/34.jpg)
Min-Cut Segmentation
cut
A
B
∑∈∈
=BjAi
ijWBAcut,
),(
Segmentation by min-cut:
Find A,B such that cut(A,B) is minimal.
(Wu and Leahy 1993)
![Page 35: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/35.jpg)
Normalized-Cut Segmentation
⎟⎟⎠
⎞⎜⎜⎝
⎛+= ∑
∈∈ )(1
)(1),(
, BvolAvolWBANcutBjAi
ij
Segmentation by normalized-cut:
Find A,B such that Ncut(A,B) is minimal.
(Shi and Malik 2000)
cut
Min-cut segmentation favors small segments.
∑∈∈
=AjAi
ijWAvol,
)(where
![Page 36: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/36.jpg)
Normalized-Cut Segmentation -Examples
(from Cohen-Or 2005)
![Page 37: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/37.jpg)
Shape Matching / Object Recognition
Model #1 Model #2 Model #3 Model #4
? ?? ?
Which Model matches the Measurement?• Which Model • What is the transformation from Model to Measurement(translation, rotation, scale,…)
![Page 38: (1 slide per page)](https://reader033.fdocuments.in/reader033/viewer/2022051506/58871b2a1a28abb3558bc5d6/html5/thumbnails/38.jpg)