1 An Implementation Sanun Srisuk of EdgeFlow
2 Outline Problem Statement Theory & Algorithm Results Conclusion
3 Problem Statement Segmentation using small scaleSegmentation using large scale
4 Theory & Algorithm is a pixel in an image. is an edge energy at location s along the orientation theta. is the probability of finding an image boundary in the direction theta from s. is the probability of finding an image boundary in the direction theta+pi from s.
5 Gaussian Function
6 3-D GD mask
7 GD mask in different theta
8 3-D DOOG mask
9 DOOG mask in different theta
10 Intensity Edges The edge energy E(s, theta) at scale sigma is defined to be the magnitude of the gradient of the smoothed image, which is obtained by smoothing the original image I(x,y) with a Gaussian kernel
11 Intensity Edges P(s, theta) is the probability of finding an image boundary in the direction theta from s.
12 Conventional Edge Detection & EdgeFlow
13 Edge Flow Vector where is a complex number with its magnitude representing the resulting edge energy and angle representing the flow direction.
14 Results red greenblue intensity
15 Intensity & Texture EdgeFlow
16 Total EdgeFlow
17 Segmentation Results
18 Segmentation Results
19 EdgeFlow
20 EdgeFlow
21 EdgeFlow Results
22 EdgeFlow Results
23 EdgeFlow Results
24 EdgeFlow Results
25 Segmentation results in different theta
26 Segmentation results in different theta
27 Conclusion EdgeFlow using a predictive coding model to identify and integrate the direction of change in image attributes such as color, texture, and phase discontinuities, at each image location.
28 The End.