-
Notifications
You must be signed in to change notification settings - Fork 36
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
- Loading branch information
Showing
8 changed files
with
283 additions
and
100 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,106 @@ | ||
| #ifndef SFERES_KDTREE_STORAGE_HPP_ | ||
| #define SFERES_KDTREE_STORAGE_HPP_ | ||
|
|
||
| #include <Eigen/Core> | ||
| #include <algorithm> | ||
| #include <numeric> | ||
| #include <vector> | ||
|
|
||
| #ifdef USE_KDTREE | ||
| #define LIBSSRCKDTREE_HAVE_BOOST | ||
|
|
||
| #include <ssrc/spatial/kd_tree.h> | ||
|
|
||
| namespace sferes { | ||
| namespace qd { | ||
| namespace container { | ||
| /// compute the k-nearest neighbors using a kd-tree (optional external library) | ||
| /// this is effective in low-dimensional spaces (< 10) | ||
| /// but not in high-dimensional spaces (> 10) | ||
| template <typename Value, int Dim> class KdtreeStorage { | ||
| public: | ||
| using point_t = std::array<float, Dim>; | ||
| using tree_t = ssrc::spatial::kd_tree<point_t, Value>; | ||
| using data_t = std::pair<Eigen::VectorXd, Value>; | ||
| using iterator_t = typename tree_t::iterator; | ||
| using const_iterator_t = typename tree_t::const_iterator; | ||
|
|
||
| /// number of elements in the archive | ||
| size_t size() const { return _data.size(); } | ||
|
|
||
| /// accesss to the elements | ||
| iterator_t begin() { return _data.begin(); } | ||
| iterator_t end() { return _data.end(); } | ||
| const_iterator_t begin() const { return _data.begin(); } | ||
| const_iterator_t end() const { return _data.end(); } | ||
|
|
||
| /// Add a point to the storage | ||
| void add(const Eigen::VectorXd& p, const Value& v) | ||
| { | ||
| point_t pp = convert_point(p); | ||
| _data.insert(pp, v); | ||
| } | ||
|
|
||
| /// Remove a point (return false if not found) | ||
| void remove(const Eigen::VectorXd& v) | ||
| { | ||
| point_t p = convert_point(v); | ||
| _data.remove(p); | ||
| } | ||
|
|
||
|
|
||
| // return the data that correspond to the k closest neighbors | ||
| // (if you neeed indices, store them in Value) | ||
| std::vector<data_t> knn(const Eigen::VectorXd& pp, int k) const | ||
| { | ||
| assert(!_data.empty()); | ||
| k = std::min(k, (int)_data.size()); | ||
| point_t p = convert_point(pp); | ||
| auto range = _data.find_nearest_neighbors(p, k, true); | ||
| std::vector<data_t> res; | ||
| for (auto it = range.first, end = range.second; | ||
| it != end && res.size() < k; ++it) | ||
| res.push_back(std::make_pair(convert_point(it->first), it->second)); | ||
| assert(res.size() <= k); | ||
| assert(res.size() != 0); | ||
| return res; | ||
| } | ||
|
|
||
| // return the data that correspond to the closest neighbor | ||
| // (if you neeed indices, store them in Value) | ||
| data_t nearest(const Eigen::VectorXd& p) const | ||
| { | ||
| assert(!_data.empty()); | ||
| point_t pp = convert_point(p); | ||
| auto it = _data.find_nearest_neighbor(pp, true); | ||
| return std::make_pair(convert_point(it->first), it->second); | ||
| } | ||
|
|
||
| /// Return true if the point p is already in the archive | ||
| bool contains(const Eigen::VectorXd& p) const | ||
| { | ||
| assert(0); // not implemented yet | ||
| } | ||
|
|
||
|
|
||
| point_t convert_point(const Eigen::VectorXd& v) const { | ||
| assert(v.size() == Dim); | ||
| point_t p; | ||
| for (size_t i = 0; i < v.size(); ++i) | ||
| p[i] = v[i]; | ||
| return p; | ||
| } | ||
| Eigen::VectorXd convert_point(const point_t& v) const { | ||
| Eigen::VectorXd p(v.size()); | ||
| for (size_t i = 0; i < v.size(); ++i) | ||
| p[i] = v[i]; | ||
| return p; | ||
| } | ||
| protected: | ||
| tree_t _data; | ||
| }; | ||
| } // namespace container | ||
| } // namespace qd | ||
| } // namespace sferes | ||
| #endif | ||
| #endif |
Oops, something went wrong.