2
votes

I'm relatively new to Matlab and I'm trying to find a part of an image inside of another image, but not just a simple cropped part that I could normally find with normxcorr2.

Specifically, I am given a cropped-out body with a black background in 2-dimensional grayscale jpg and I have to find where it comes from in the source image (also a 2-d gray jpg).

I am required to use segmentation and thresholding to obtain the "important features" of the cropped out body, which I have done and this is the result (with default thresholding levels), which saves in the exact same format as before.

segmented image

What I have tried: The version of Matlab I have available to me only includes the Image Processing Toolbox, and that is all I am allowed to use in terms of toolboxes (no Computer Vision Toolbox). I have tried is using OpenSURF for feature matching, and this was the result:

result image

The result is less than favourable. Is there any other way I can accomplish this? Please keep in mind that I am new to image processing and Matlab.

1

1 Answers

1
votes

If you have the Computer Vision System Toolbox, please try the Object Detection in a Cluttered Scene Using Point Feature Matching example.