Smoothing the edges of an image

Question

David Dijemeni 2011-12-4

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/23006-smoothing-the-edges-of-an-image

I have detected different parts of an MRI scan but the edges of the detected part is very rough in some places. How can I smoothen the edges of the detected part?

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

David Young 2011-12-4

3
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/23006-smoothing-the-edges-of-an-image#answer_30232

If you use the edge() function, try the 'canny' option and adjust the sigma parameter to give edges as smooth as you need them.

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

David Dijemeni 2011-12-4

This gives me the result I wanted to a good approximation. Thanks David!

请先登录，再进行评论。

Answer 2

Image Analyst 2011-12-4

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/23006-smoothing-the-edges-of-an-image#answer_30257

You can smooth a list of coordinates by using conv(xCoords, ones(1,WindowSize)/windowSize) - similar for the y coordinates. You just average together a section of the curve. You need to do a little fixing at the beginning and end of the array to handle wrap around, but that's not hard. This will give you smoothed shaped - the bigger the blurring window the smoother the shape is.

I'm using this method now in one project. Actually I wanted to smooth inward incursions only, not outward protuberances so I OR'ed with the original (gotten by using poly2mask).

There are other ways to smooth it, such as by using imclose, blurring your image before edge detection, etc. Results depend on your image so you might experiment with different methods to see which one delivers the appearance you're seeking.