README:

THIS IS A PROJECT DONE AS A PART OF THE "CIS-581: COMPUTER VISION AND COMPUTATIONAL PHOTOGRAPHY" COURSE AT UNIVERSITY OF PENNSYLVANIA, TAUGHT BY PROF.JIANBO SHI. 


-> RUN THE PROGRAM FROM "wrapper.py"
-> THE THREE RUNNING CASES ARE GIVEN, WITH ONE ACTIVE AND TWO INACTIVE. TO RUN ANY OTHER CASE, COMMENT OUT THE ACTIVE CASE AND UNCOMMENT THE CASE YOU WANT TO RUN

-> FOR THE TEST IMAGE CASE, THE IMAGES LOADED ARE "1_M.jpg" AND "1_R.jpg". TO RUN THE OTHER COMBINATION,  CHANGE M TO L AND R TO M.

-> BOTH SETS 1 AND 2 RUN ON THE PARAMETERS THAT ARE DEFINED THROUGHOUT, WHILE THE PARAMETERS CHANGE FOR THE TEST IMAGES.

-> FOR THE TEST IMAGES, THE FEATURES TO BE EXTRACTED ARE FROM SHI&TOMASI, AND NOT HARRIS. THIS LINE IS COMMENTED OUT IN "corner_detector.py"

-> FOR THE TEST IMAGES, THE VALUE OF THE RATIO TEST IS 0.95, AND NOT 0.9. THIS LINE IS COMMENTED OUT IN "feat_match.py"

-> FOR THE TEST IMAGES, THE NUMBER OF ITERATIONS IS 10000 AND NOT 1000. THIS LINE IS COMMENTED OUT IN "ransac_est_homography.py"

-> IN ALL CASES, THE DESCRIPTOR IS OBTAINED USING A FINAL MAX POOLING, RATHER THAN TAKING IMAGES AT INTERVALS OF 5 PIXELS.

-> FOR EVERY CASE, THE FIGURES THAT WILL BE DISPLAYED ARE THE ANMS POINTS, THE MATCHING AFTER RANSAC, AND THE MOSAIC. THIS IS DONE FOR EVER PAIR OF IMAGES, UNTIL THE FINAL MOSAIC.