![Page 1: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/1.jpg)
Filters and Edges
![Page 2: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/2.jpg)
![Page 3: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/3.jpg)
![Page 4: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/4.jpg)
Zebra convolved with Leopard
![Page 5: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/5.jpg)
![Page 6: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/6.jpg)
![Page 7: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/7.jpg)
![Page 8: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/8.jpg)
![Page 9: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/9.jpg)
![Page 10: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/10.jpg)
![Page 11: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/11.jpg)
![Page 12: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/12.jpg)
![Page 13: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/13.jpg)
![Page 14: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/14.jpg)
![Page 15: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/15.jpg)
![Page 16: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/16.jpg)
![Page 17: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/17.jpg)
![Page 18: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/18.jpg)
![Page 19: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/19.jpg)
![Page 20: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/20.jpg)
![Page 21: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/21.jpg)
50 100 150 200 250 300 350 400 450 500
50
100
150
200
250
300
350
400
450
500
50 100 150 200 250
50
100
150
200
250
![Page 22: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/22.jpg)
20 40 60 80 100 120
20
40
60
80
100
120
20 40 60 80 100 120
20
40
60
80
100
120
10 20 30 40 50 60
10
20
30
40
50
60
10 20 30 40 50 60
10
20
30
40
50
60
![Page 23: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/23.jpg)
5 10 15 20 25 30
5
10
15
20
25
30
5 10 15 20 25 30
5
10
15
20
25
30
2 4 6 8 10 12 14 16
2
4
6
8
10
12
14
16
2 4 6 8 10 12 14 16
2
4
6
8
10
12
14
16
![Page 24: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/24.jpg)
![Page 25: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/25.jpg)
Edge Detection as a signal detection problem
Goal: find meaningful intensity boundaries.
![Page 26: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/26.jpg)
Simplest Model: (Canny)
Edge(x) = a U(x) + n(x)
U(x)
x=0
Convolve image with U and find points with high magnitude. Choose value by comparing with a threshold determined by the noise model.
?
![Page 27: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/27.jpg)
Probability of a filter response on an edge
Probability of a filter response off an edge
![Page 28: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/28.jpg)
Fundamental limits on edge detection
![Page 29: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/29.jpg)
Need to take into account base edge rate.
![Page 30: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/30.jpg)
Smoothing and Differentiation• Issue: noise
– smooth before differentiation– two convolutions to smooth, then differentiate?– actually, no - we can use a derivative of
Gaussian filter• because differentiation is convolution, and
convolution is associative
![Page 31: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/31.jpg)
There are three major issues: 1) The gradient magnitude at different scales is different; which should we choose? 2) The gradient magnitude is large along thick trail; how do we identify the significant points? 3) How do we link the relevant points up into curves?
![Page 32: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/32.jpg)
The scale of the smoothing filter affects derivative estimates, and alsothe semantics of the edges recovered.
1 pixel 3 pixels 7 pixels
![Page 33: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/33.jpg)
Computing Edges via the gradient
![Page 34: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/34.jpg)
We wish to mark points along the curve where the magnitude is biggest.We can do this by looking for a maximum along a slice normal to the curve(non-maximum suppression). These points should form a curve. There arethen two algorithmic issues: at which point is the maximum, and where is thenext one?
![Page 35: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/35.jpg)
Non-maximumsuppression
At q, we have a maximum if the value is larger than those at both p and at r. Interpolate to get these values.
![Page 36: Filters and Edges. Zebra convolved with Leopard](https://reader031.vdocuments.us/reader031/viewer/2022020117/56649d605503460f94a41857/html5/thumbnails/36.jpg)
Predictingthe nextedge point
Assume the marked point is an edge point. Then we construct the tangent to the edge curve (which is normal to the gradient at that point) and use this to predict the next points (here either r or s).