Re: Help watershed oversegmentation

Dear Tilman JOCHEMS,

I vote for your QLisp image processing program! Simple & Effective!
It's similar to Matlab but lightweight and more comprehensive
commands. I like it.

I've tested your qLisp code on some teeth images. The result is good
for some images and under-segmentation for some other ones. Anyway,
the idea inside your code really make interested. As you know, X-ray
film contains some over-bright regions which are not teeth. Sometime
gum and teeth can't be separated.

Your method is senvitive to bright artifacts. If there are some
artificial symbols, text or logo,..., on the film, markers will be
absorbed by those artifacts.

In conclusion, I like the way the gum marker is built.

I posted a .rar file contains some processed images.
Link: http://files.filefront.com/Picturesrar/;7621008;/fileinfo.html

Thank you for all. Wait for your reply.

Phong

.

Relevant Pages
Loading