How can I use the SIFT descriptor key points I get from an image and run RANSAC (estimateGeometricTransform) on it?

Question

Hi everyone I am relatively new to programming in general and MatLab. I have an assignment that wants me to compare different MatLab detection functions (SURF, BRISK, FAST and etc,) with the SIFT descriptor.

I have the code for MatLab functions and checking the key point on that and then running the estimateGeometricTransform which is sort of similar to RANSAC and check the results.

What I am having difficulties with is finding a SIFT code using vlfeat to extract the key points and then run estimateGeometricTransform on them. All I can find online of ready codes are the SIFT RANSAC for image mosaics and stitching, which is not what I need. I need to use the same image, create some sort of rotation or scale differences and see how the keypoints will match for different kind of images.

All the codes online seem to be too difficult to handle. Can someone guide me to where I can find what I am looking for, or how I can implement it if it is even possible?

Answer 1

1. Extract keypoints from SIFT from vl_sift
2. Find the coordinates point or matching point of image. Matching point can be run by vl_ubcmatch.
3. run estimateGeometricTransform with the keypoints.

Example: [f1,d1] = vl_sift(I1);

In f1 or frame 1, in the first row which is the x coordinates and second row is the y coordinates.