HOW TO SEGMENT FUSED CHARACTERS ?

1 次查看（过去 30 天）
soumyadeep
soumyadeep 2014-4-4
评论： Image Analyst 2014-4-4
hello everyone!! i am trying to segment touched characters. for example
the above figure shows three characters fused into one. i tried the watershed algorithm and it segmented the above figure into many segments.
which as you can see divides the characters into three parts along with many more segmentation. but the problem is after dilating it it again fuses into one. so can anyone help me with this. or suggest some other method of " SEGMENTING FUSED CHARACTERS "??
btw the code for watersegment which i am using is
j= imread('image 1.png'); d=-bwdist(~j); l=watershed(d); j(l==0)=0; imshow(j);

请先登录，再回答此问题。

回答（1 个）

Image Analyst
Image Analyst 2014-4-4
The image is very low resolution which makes it hard. Plus I'm not even sure where to split apart the blob even if I were asked to do it manually. It's just not clear. I have no idea how many characters are there and what they might be. Could be anywhere from 3 to 6 but I don't recognize any of them.
  2 个评论
soumyadeep
soumyadeep 2014-4-4
hello sir! well doing it manually is not an option because there are like 50 data sets we need to segment and to segment each sample manually would be very tedious job. well in the above example the image should be segmented in following characters. well im using photoshop to crop the characters here.
btw will it help if we use a high resolution image?? thank you!!
Image Analyst
Image Analyst 2014-4-4
I would not know how to divide up that assortment of blobs into the 3 new assortment of blobs, unless you just want to take the left third, middle third, and right third of the image and see which centroids like in those zones. The horizontal location of the blobs is about the only thing that I see that predictably distinguishes one blob from another.

请先登录，再进行评论。

类别

在 Help Center 和 File Exchange 中查找有关 Image Segmentation and Analysis 的更多信息

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by