I use this code:
int c_count =cv::flann::hierarchicalClustering<L2<float> >(data,cluster,kmeans_param);
I got this error:
error C2146: syntax error : missing ';' before identifier 'ElementType'
How can I solve it?
My features is 3d (x,y,theta) that I want to cluster these. when the classification was done.How can I access to cluster id for each feature?
I solve it by below code.
Mat1f cluster_data(15,3);
cluster_data(0,0) = 1; cluster_data(0,1) = 1;cluster_data(0,2) = 1;
cluster_data(1,0) = 2; cluster_data(1,1) = 2;cluster_data(1,2) = 1;
cluster_data(2,0) = 2; cluster_data(2,1) = 1;cluster_data(2,2) = 1;
cluster_data(3,0) = 1; cluster_data(3,1) = 1;cluster_data(3,2) = 2;
cluster_data(4,0) = 1; cluster_data(4,1) = 2;cluster_data(4,2) = 2;
cluster_data(5,0) = 100; cluster_data(5,1) = 99;cluster_data(5,2) = 98;
cluster_data(6,0) = 101; cluster_data(6,1) = 100;cluster_data(6,2) = 99;
cluster_data(7,0) = 103; cluster_data(7,1) = 90;cluster_data(7,2) = 97;
cluster_data(8,0) = 99; cluster_data(8,1) = 98;cluster_data(8,2) = 102;
cluster_data(9,0) = 98; cluster_data(9,1) = 100;cluster_data(9,2) = 102;
cluster_data(10,0) = 1001; cluster_data(10,1) = 997;cluster_data(10,2) = 996;
cluster_data(11,0) = 1002; cluster_data(11,1) = 1001;cluster_data(11,2) = 999;
cluster_data(12,0) = 1000; cluster_data(12,1) = 999;cluster_data(12,2) = 998;
cluster_data(13,0) = 999; cluster_data(13,1) = 998;cluster_data(13,2) = 1000;
cluster_data(14,0) = 998; cluster_data(14,1) = 1003;cluster_data(14,2) = 1003;
cvflann::KMeansIndexParams k_params(10, 1000, cvflann::FLANN_CENTERS_KMEANSPP,0.01);
Mat1f centers(10,3);
centers.setTo(0);
int count = cv::flann::hierarchicalClustering<cv::flann::L2<float> >(cluster_data,centers,k_params);
cout << cluster_data << endl;
cout << centers << endl;
Asked: 2013-07-30 16:27:48 -0600
Seen: 2,557 times
Last updated: Jul 31 '13
The second post is not related to first post.The current post correspond to code & other post correspond to comparison of the algorithms. Please respect the right of other forum members .Please Open the current post and other post back to previous state.
I do not think I didn't respect your right as forum member, I merely merged two topics that handle the exact same topic, being different clustering types. Discussing a technique and then trying to solve problems with using that specific algorithm seems reasonably the same to me. Topics restored like you asked ... will let others deal with it if they have problems.
Also, guidelines say to update your original topic, when you have find a solution which yields new problems or warnings.
Thank you.