Segmenting using X-Y co-ordinate axis

Hey guys !

I want to know how to segment an image using the X-Y co-ordinate axis. I am trying to make an Optical Braille Recognition software and I want to use X-Y frame to segment individual Braille characters. I know the exact dimensions of a standard Braille character on the co-ordinate frame. How do I go about it ?

Answer 1

[Walter Roberson]

If you know the exact dimensions on a coordinate frame, then you would not do "segmentation", just object location: once you know the object location, you would just snip out the relevant portion of the array (or label that portion of the array all with the same label.)

How regular is the placement of the braille characters? For example, if you know the coordinates of one character then by knowing the standard dimensions would you then be able to trivially calculate the location of the next character on the row? Are the columns also regularly spaced?

This also raises to me the possibility that even if the grid of braille characters is exactly predictable, that the image might be titled with respect to the grid. Is that an aspect that you need to be concerned about? If so then you would probably need to start with some code to determine the image angle and then rotate the image back in to alignment using imrotate.

[Siddharth Mallya]

Well, u r right. There is always an the problem of tilted axis. But, for the purpose of this program, we are just testing if OBR using MATLAB can be done for Braille at all - so we are using printed braille (using Braille font, i.e.)

U r also right about the object location and snipping. I know the exact location and dimensions of each character. How would u suggest I go about the programming tho ? I usually never ask for help in forums for programming but being forced to take up MATLAB has reduced me to this .. :(

[Walter Roberson]

How would you do it in a different programming language? Your need is to get it working first; _then_ you can worry about making it more MATLAB-like.

I suggest you start by thresholding the image. Then find the first row that has content -- the first row that has ANY non-zero content (that was a hint.) You can do the same thing for columns. Statistically, the intersection should be the upper-left corner of the upper-left braille cell, but only statistically: you need to cross-check against the possibility that the upper row contains only periods, as in that case the upper edge would be a line lower than otherwise.

[Siddharth Mallya]

Okay .. Right now, I have managed to reduce the noise of the pixel, and using bwlabel and regionprops find the braille dots on the document image. if u use a standard a4 size paper, the starting point is always the same. so i am moving the origin to that (x,y) space. further, i am thinking ill segment using braille cell size.

is the approach ok?

[Walter Roberson]

Perhaps the position is close enough to being the same with laser printers, but with dot matrix printers the feed mechanisms were seldom reliable, so the position you determined on one A4 page would only be a close approximation but not exact. Best to take a couple of steps to re-register the first cell.

[Siddharth Mallya]

Wow .. I hadn't thought of that. Thanks for the heads up !

Answer 2

[Brett Shoelson]

XY coordinates are non-standard (in MATLAB) for displaying images. But non-standard doesn't mean "not doable." Using the AXIS XY command will set the coordinate system origin to be in the lower left corner. The x-axis will be horizontal with values increasing from left to right, and the y-axis will be vertical with values increasing from bottom to top. However, this will also flip your image, so you may have to account for that:

img = imread('cameraman.tif');
img = flipud(img);
imshow(img)
axis xy
impixelinfo

Cheers,

Brett

[Siddharth Mallya]

@Brett,

Thank you for your quick reply. If this method is "non-standard", what is considered as the standard for my kind of problem statement ? I am a complete newbie to MATLAB and would love all the help I could get.

Sid

[Brett Shoelson]

Welcome to MATLAB, Sid! Take a look at the different options for AXIS in our doc. Then display an image and see what happens when you apply those options.

Cheers,

Brett