suimo comments

Results 11 comments of


                                            suimo

performance on low dimensional datasets

Thanks a lot for your reply! Could you give me some advice that which algorithms are better for low dimensional dataset.

performance on low dimensional datasets

Thanks for your reply. I will try it.

BA failed

@kiyology thanks for your reply. I have tried to reinstall "libsuitesparse-dev",but it seems not work,the error still exists. move over,I have some other confusions: 1. I found that it use...

performance on low dimensional datasets

Hi, Thanks for your reply! I use KDTreeSingleIndex in flann. May be I didn't use the OpenMP to speed up. Could it be the reason that hnsw slower than flann?

performance on low dimensional datasets

I tested both algorithm in a single thread. And the dataset is collected from lidar sensor

performance on low dimensional datasets

By the way, I used a for loop to input the query point one by one.

performance on low dimensional datasets

Hi, @yurymalkov Thanks for your detailed reply. I just used original flann(https://github.com/mariusmuja/flann). @searchivarius the dataset is collected from velodyne lidar(VP-16). There are some lidar points dataset on the KITTI Benchmark,...

performance on low dimensional datasets

I just used the definition "#pragma omp parallel for" when building index, if not, the building time will be 12364.4 ms

performance on low dimensional datasets

So it's better to use tree-based algorithm for low dimensional dataset, right?

performance on low dimensional datasets

I got it.Thanks for your advice!