Graph Extraction from Handwritten Character.??

1 回表示 (過去 30 日間)
Mohammad Bhat
Mohammad Bhat 2017 年 6 月 27 日
コメント済み: Walter Roberson 2017 年 6 月 27 日
Hi,
I have extracted important nodes from handwritten character. After that I have to connect those nodes by "Neighbouhood analysis." How can I do that? Any ideas or any suggestions??
The image is attached below.
  2 件のコメント
Image Analyst
Image Analyst 2017 年 6 月 27 日
編集済み: Image Analyst 2017 年 6 月 27 日
Not sure what that means. It sounds like you got that phrase from some paper. If so, doesn't the paper describe how to do it?
You could threshold and then call imclose() and then bwmorph(BW, 'skel', inf).
Mohammad Bhat
Mohammad Bhat 2017 年 6 月 27 日
No sir, I got important points from character , these points are choosen such that inherent structure of the character image is preserved....After getting these points I have to join them such that underlying graph is formed. How can I do that..?

サインインしてコメントする。

回答 (1 件)

Walter Roberson
Walter Roberson 2017 年 6 月 27 日
wb = im2bw( imread('untitled.jpg') );
bw = ~wb;
se = strel('disk',4);
joined = imclose(bw, se);
thin = bwmorph(joined, 'thin', inf);
imshow(thin)
  2 件のコメント
Mohammad Bhat
Mohammad Bhat 2017 年 6 月 27 日
Nothing, other than white dot in a black backgrond
Walter Roberson
Walter Roberson 2017 年 6 月 27 日
The above is the output. You can see that all of the neighbours are joined together, and that the result is one continuous outline for the character, together with a dot for the black spot to the left in the original (you did not say to filter that out.) The nodes have been connected using neighborhood analysis -- imclose() is morphological neighborhood processing that does dilation and then erosion.

サインインしてコメントする。

Community Treasure Hunt

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

Start Hunting!

Translated by