you can run the main.py file to find some test results This repo was developed as a recreation of methods layed by a paper by David G. Lowe you can find [here]{https://www.cs.ubc.ca/~lowe/papers/ijcv04.pdf}.
- involves taking sparse sift keypioints over the image
- Use a brute force feature matcher to find good matches by using knn taking M and comparing the distance of it too the second good match - This is done by comapring the distance of a first match to a second match by a distance of .75
- this step involves taking a cluster hough transform from the identified.
- from the following data the algorithim writes abounding box of various size and orientation
