matlab - Character extraction - dots are recognized seperately from the character -

matlab - Character extraction - dots are recognized seperately from the character -

i'm doing character recognition regional language. while extracting image, dots beingness separately identified characters. please help me .

%% plot bounding box n=1:size(propied,1) rectangle('position',propied(n).boundingbox,'edgecolor','g','linewidth',2) end hold off %% characters beingness extracted figure n=1:ne [r,c] = find(l==n); n1=imagen(min(r):max(r),min(c):max(c)); imshow(~n1); end

please help

original code: http://www.mathworks.com/matlabcentral/fileexchange/22922-image-segmentation-extraction

since doing character/text recognition, more want collections of words or lines of text, , not individual characters. , if want latter, more robust after have identified individual words.

so, simplest approach here using standard morphological opening (assuming text black, otherwise utilize closing) operator. start big horizontal structuring element (se). applying opening se split image in lines of text. in each line utilize shorter horizontal se obtain individual words. each word consider vertical se opening such joins accents , other typographical details.

for example, here input image, opening horizontal se of radius 35, opening horizontal se of radius 7, , opening vertical se of radius 7.

i didn't apply 3rd operation in isolated components, should not risk joining 2 lines of text. , assuming straight horizontal lines of text, of course. drawing bounding boxes on final image gives result after:

note letters ("ty", , "ny") connected in beginning, appear single letter in output. separate problem handled, might or not concern you.

matlab image-processing