Wednesday, 5 February 2014

Image matching in MATLAB

Image matching- which means comparing two images for identical features. Often its practical usages are biometric applications such as finger print identification, Iris matching etc. Two images matches when their features coincides and these feature could be edges, corners, blobs, color, shape etc. By identifying these specific features in an image and comparing it with the features of other image, we can determine whether the images matches or not. Below shown is a demonstration of the image matching in the application of finger print recognition in MATLAB,

In this program, the number of edges in the image1 is compared with that of the number of edges in the image2. If the image matches, the GUI returns a 'Match' dialogue box.


1. Load the image into MATLAB
2. Convert to grayscale image
3. Apply the edge detector of your choice after removing noises, if any.
4. Create a copy of this image for comparison (say img1 and img2)
5. Now traverse through the every pixel of both the images using a looping statement,
6. And, check the pixel value at the corresponding pixel
7. If the comparison percentage if greater than certain percentage( say 90, as per your wish), then the finger prints matches, else, the finger prints are not a match

Note: This not the exact way how the conventional biometric scanners work, but, this method can be used for rough matching purpose

1 comment:

  1. please give me the code at