Image Segmentation I

1
Image Segmentation I

• Comp344 Tutorial
• Kai Zhang

2
Outline

• Line Edge detection
• Hough transform

3
Line Detection

• Apply a mask and compute the image response
• Design of Mask determines what kind of line is
  detected
• -1 -1 -1 -1 -1 2 -1 2 -1 2
  -1 -1
• 2 2 2 -1 2 -1 -1 2 -1
  -1 2 -1
• -1 -1 -1 2 -1 -1 -1 2 -1 -1
  -1 2
• Horizontal 45o vertical
  -45o

4
(No Transcript)
5
Codes

• f imread('wirebond.tif')
• w 2 -1 -1 -1 2 -1 -1 -1 2
• g imfilter(double(f),w)
• imshow(g,)
• gtop g(1120,1120)
• gtop pixeldup(gtop, 4)
• figure,imshow(gtop, )
• gbot g(end - 119end, end-119end)
• gbot pixeldup(gbot, 4)
• figure,imshow(gbot, )
• g abs(g)
• figure,imshow(g,)
• T max(g())
• g g gt T
• figure,imshow(g,)

6
Hough Transform

• A method for line detection on binary images
• Transform points in the x-y plane into the
  parameter space (\rho, \theta)
• Detect local maximum on the parameter space (kind
  of voting), which corresponds to strong line
  signals
• Complexity much lower than naïve method

7
Sparse matrices in matlab

• Sparse matrix Matrices that contain only a small
  number of non-zero elements
• Sparse matrix can reduce both the storage and the
  computational complexity
• Example
• A 0 0 0 5 0 2 0 0 1 3 0 0 0 0 4 0
• S sparse(A)
• S
• (3,1) 1
• (2,2) 2
• (3,2) 3
• (4,3) 4
• (1,4) 5

8
Example 1
9
Codes

• f zeros(101,101)
• f(1,1) 1 f(101,1) 1 f(1,101) 1
• f(101,101) 1 f(51,51) 1
• figure,imshow(f,)
• H hough(f)
• figure, imshow(histeq(H),)
• H, theta, rho hough(f)
• figure, imshow(theta, rho, histeq(H), ,
  'notruesize')
• axis on axis normal
• xlabel('theta'),ylabel('\rho')

10
Example 2
11
codes

• f double(imread('building.tif'))
• canny,tc edge(f,'canny',0.04 0.1, 1.5)
• figure,imshow(canny,)
• f g_canny_default
• H, theta, rho hough(f,0.5)
• figure, imshow(theta, rho, H/30,
  'notruesize'),axis on axis normal
• r,c houghpeaks(H, 5)
• hold on plot(theta(c),rho(r),'linestyle','none','
  marker','s','color','w')
• lines houghlines(f, theta, rho, r,c)
• figure,imshow(f,)hold on
• for k 1length(lines)
• xy lines(k).point1 lines(k).point2
• plot(xy(,2),xy(,1),'LineWidth',4,'Color',.9 .0
  .0)
• end