Using OpenCV to compute the distances between a vector of points from a given point.
e.g.
std::vector<cv::Point2f> pointPositions
cv::Point2f center
for(const auto& pt : pointPositions
auto d = cv::norm(center-pt);
What are some more efficient & alternative ways to do this for a "sufficiently large data set" ? What would be a suitable multi-threaded approach to compute distances which scales well across multiple processing cores (close to or nearly 4x improvement when using 4 cores instead of 1) ?
p.s. I can avoid computing the square root and keep the distance^2