I am looking for a freelancer who has very deep understanding of Computer Vision and Machine Learning who can develop a robust image matching algorithm. For the project, I have tried feature based/keypoint matching algorithm including ORB, SIFT, Akaze, Brisk and others along with homography implementation; yet I could not receive any satisfactory results. Use the given query images in the query folder to accurately match the target images. You can use either classical computer vision approaches or unsupervised learning to obtain minimum 95% accuracy. Supervised learning may not be suitable for this task since new target images will be generated by the user.
NOTE: The deliverable will be only acceptable for payment if your algorithm meets the performance target on an evaluation dataset separate from the test data given to you. I would advise doing a preliminary work on the sample dataset before submitting an offer.
About the recuiterMember since Mar 14, 2020 L Ben
from Rangpur, Bangladesh