Question

as i know knn find smallest distance from all samples.it means it is the nearest sample. but what if i want distance of my new data from all samples? how can I get for example 3 or 4 nearest samples to my new data?
Thank you

Answer 1

K nearest neighbour? That means that you can find K nearest samples. You should read documentation carefully.
Example:

    kdtree.knnSearch(((cv::Mat)objects_coordinates[i_pts]).t(),
        indices, dists, num_of_neighbours, cv::flann::SearchParams(32));