charecter segmentation matlab

we will use the above image for processing

1.add the image to matlab work space

ch=imread('char.jpg');

2.convert the RGB image to gray scale using rgb2gray function
Gch=rgb2gray(ch);

3.convert this gray scale image to BW comparing with some threshold
bw= im2bw(Gch,graythresh(Gch));

now u will get a binary image containing 1 r 0 's which will be easy to process

4. find the edge's in that binary image
Ibw= edge(uint8(bw));
imshow(Ibw);

5.using Morphological process to enhance the image
se = strel('disk',4);
bw2 = imdilate(Ibw, se);
imshow(bw2);



6. filling the holes in the dilated image
fill= imfill(bw2,'holes');
imshow(fill)
7. labeling the parts of the image and marking their position
[lab n] = bwlabel(fill);
props = regionprops(lab);
Z = [props.BoundingBox];
Z= reshape(Z,[4 n]);


8. display the results
imshow(I);
hold on;
for i = 1:n
rectangle('position',z(:,i),'edgecolor','r');
end


this can be used with a neural network to recognize patterns
we will see how to do this in next blog