Hi:
here is a method to connect line segment:
distance transform --> use the max distance you want to connect as threshold --> do watershed

But I think Canny is not a good choice to extract region(not continue and too many noise),.
your image is hard, some work you can try:
1. use a snake curve to fit step by step
2. use ml function (which need many images to train)
3. if you need not full-automatic, you can use Opencv's GrabCut.

Some images attached processed by ImagePy, you can get ImagePy here:
https://github.com/Image-Py/imagepy

Best

----- 原始邮件 -----
发件人：Randy Heiland <randy.heiland@gmail.com>
收件人：scikit-image@python.org
主题：[scikit-image] shrink-wrapping filopodia
日期：2018年01月05日 01点22分

Hello, and thanks for the great software!

I thought I'd ask this community for suggestions on extracting the outline/shape of some images involving biological filopodia. I've attached a sample image where I manually captured one potential outline. And I've put my initial scripts, data, results here:
https://github.com/rheiland/image_proc

I tried to "close" (edge2.py) the noisy edges, thinking that would be a logical first step, but it didn't really do what I'd hoped.

thanks, Randy

_______________________________________________
scikit-image mailing list
scikit-image@python.org
https://mail.python.org/mailman/listinfo/scikit-image