How to establish image feature library with OpenCV?
I want to use sift features which extract from about 1000 images to establish an image feature library. How can I do these? In order to faster the progress of image matching, my idea is to store these features into a database(e.g. mysql), and then read these features from this database. Is this idea feasible? Help someone can help me.
Yes it should be possible. For the SIFT part, you can look at the corresponding tutorials.
this sounds like a bad idea. you'll probably end up using BagOfWords with some meachinelearning(like an SVM) so you need to store a BOW vocabulary (large) and a trained ml model(also large), not single sift features.
Thanks for all the advice. But I am still confused that if I have a new image and I want to find the matched images from lots of images(e.g. 1000 images) . How can I complete this work?