// Copyright (C) 2010 Davis E. King (davis@dlib.net) // License: Boost Software License See LICENSE.txt for the full license. #ifndef DLIB_GRAPH_CrEATION_H__ #define DLIB_GRAPH_CrEATION_H__ #include "graph_creation_abstract.h" #include <limits> #include <vector> #include "../string.h" #include "../rand.h" #include <algorithm> #include "sample_pair.h" namespace dlib { // ---------------------------------------------------------------------------------------- namespace impl { template <typename iterator> iterator iterator_of_worst ( iterator begin, const iterator& end ) /*! ensures - returns an iterator that points to the element in the given range that has the biggest distance !*/ { float dist = begin->distance(); iterator worst = begin; for (; begin != end; ++begin) { if (begin->distance() > dist) { dist = begin->distance(); worst = begin; } } return worst; } } // ---------------------------------------------------------------------------------------- template < typename vector_type, typename distance_function_type, typename alloc, typename T > void find_percent_shortest_edges_randomly ( const vector_type& samples, const distance_function_type& dist_funct, const double percent, const unsigned long num, const T& random_seed, std::vector<sample_pair, alloc>& out ) { // make sure requires clause is not broken DLIB_ASSERT( 0 < percent && percent <= 1 && num > 0, "\t void find_percent_shortest_edges_randomly()" << "\n\t Invalid inputs were given to this function." << "\n\t samples.size(): " << samples.size() << "\n\t percent: " << percent << "\n\t num: " << num ); out.clear(); if (samples.size() <= 1) { return; } std::vector<sample_pair, alloc> edges; edges.reserve(num); dlib::rand::kernel_1a rnd; rnd.set_seed(cast_to_string(random_seed)); // randomly sample a bunch of edges for (unsigned long i = 0; i < num; ++i) { const unsigned long idx1 = rnd.get_random_32bit_number()%samples.size(); const unsigned long idx2 = rnd.get_random_32bit_number()%samples.size(); if (idx1 != idx2) { const float dist = dist_funct(samples[idx1], samples[idx2]); if (dist < std::numeric_limits<float>::infinity()) { edges.push_back(sample_pair(idx1, idx2, dist)); } } } // now put edges into out while avoiding duplicates if (edges.size() > 0) { // sort the edges so that duplicate edges will be adjacent std::sort(edges.begin(), edges.end(), &order_by_index); out.reserve(edges.size()); out.push_back(edges[0]); for (unsigned long i = 1; i < edges.size(); ++i) { if (edges[i] != edges[i-1]) { out.push_back(edges[i]); } } // now sort all the edges by distance and take the percent with the smallest distance std::sort(out.begin(), out.end(), &order_by_distance); out.swap(edges); const unsigned long out_size = std::min<unsigned long>((unsigned long)(num*percent), edges.size()); out.assign(edges.begin(), edges.begin() + out_size); } } // ---------------------------------------------------------------------------------------- namespace impl2 { struct helper { /* This is like the sample_pair but lets the edges be directional */ helper( unsigned long idx1, unsigned long idx2, float dist ) : index1(idx1), index2(idx2), distance(dist) {} unsigned long index1; unsigned long index2; float distance; }; inline bool order_by_index ( const helper& a, const helper& b ) { return a.index1 < b.index1 || (a.index1 == b.index1 && a.index2 < b.index2); } inline bool total_order_by_distance ( const helper& a, const helper& b ) { return a.distance < b.distance || (a.distance == b.distance && order_by_index(a,b)); } } // ---------------------------------------------------------------------------------------- template < typename vector_type, typename distance_function_type, typename alloc, typename T > void find_approximate_k_nearest_neighbors ( const vector_type& samples, const distance_function_type& dist_funct, const unsigned long k, unsigned long num, const T& random_seed, std::vector<sample_pair, alloc>& out ) { // make sure requires clause is not broken DLIB_ASSERT( num > 0 && k > 0, "\t void find_approximate_k_nearest_neighbors()" << "\n\t Invalid inputs were given to this function." << "\n\t samples.size(): " << samples.size() << "\n\t k: " << k << "\n\t num: " << num ); out.clear(); if (samples.size() <= 1) { return; } // we add each edge twice in the following loop. So multiply num by 2 to account for that. num *= 2; std::vector<impl2::helper> edges; edges.reserve(num); std::vector<sample_pair, alloc> temp; temp.reserve(num); dlib::rand::kernel_1a rnd; rnd.set_seed(cast_to_string(random_seed)); // randomly sample a bunch of edges for (unsigned long i = 0; i < num; ++i) { const unsigned long idx1 = rnd.get_random_32bit_number()%samples.size(); const unsigned long idx2 = rnd.get_random_32bit_number()%samples.size(); if (idx1 != idx2) { const float dist = dist_funct(samples[idx1], samples[idx2]); if (dist < std::numeric_limits<float>::infinity()) { edges.push_back(impl2::helper(idx1, idx2, dist)); edges.push_back(impl2::helper(idx2, idx1, dist)); } } } std::sort(edges.begin(), edges.end(), &impl2::order_by_index); std::vector<impl2::helper>::iterator beg, itr; // now copy edges into temp when they aren't duplicates and also only move in the k shortest for // each index. itr = edges.begin(); while (itr != edges.end()) { // first find the bounding range for all the edges connected to node itr->index1 beg = itr; while (itr != edges.end() && itr->index1 == beg->index1) ++itr; // If the node has more than k edges then sort them by distance so that // we will end up with the k best. if (static_cast<unsigned long>(itr - beg) > k) { std::sort(beg, itr, &impl2::total_order_by_distance); } // take the k best unique edges from the range [beg,itr) temp.push_back(sample_pair(beg->index1, beg->index2, beg->distance)); unsigned long prev_index2 = beg->index2; ++beg; unsigned long count = 1; for (; beg != itr && count < k; ++beg) { if (beg->index2 != prev_index2) { temp.push_back(sample_pair(beg->index1, beg->index2, beg->distance)); ++count; } prev_index2 = beg->index2; } } // now sort temp so that we can avoid duplicates in the final loop below std::sort(temp.begin(), temp.end(), &order_by_index); // now put edges into out while avoiding duplicates if (temp.size() > 0) { out.reserve(temp.size()); out.push_back(temp[0]); for (unsigned long i = 1; i < temp.size(); ++i) { if (temp[i] != temp[i-1]) { out.push_back(temp[i]); } } } } // ---------------------------------------------------------------------------------------- template < typename vector_type, typename distance_function_type, typename alloc > void find_k_nearest_neighbors ( const vector_type& samples, const distance_function_type& dist_funct, const unsigned long k, std::vector<sample_pair, alloc>& out ) { // make sure requires clause is not broken DLIB_ASSERT(k > 0, "\t void find_k_nearest_neighbors()" << "\n\t Invalid inputs were given to this function." << "\n\t samples.size(): " << samples.size() << "\n\t k: " << k ); out.clear(); if (samples.size() <= 1) { return; } using namespace impl; std::vector<sample_pair> edges; // Initialize all the edges to an edge with an invalid index edges.resize(samples.size()*k, sample_pair(samples.size(),samples.size(),std::numeric_limits<float>::infinity())); // Hold the length for the longest edge for each node. Initially they are all infinity. std::vector<float> worst_dists(samples.size(), std::numeric_limits<float>::infinity()); std::vector<sample_pair>::iterator begin_i, end_i, begin_j, end_j, itr; begin_i = edges.begin(); end_i = begin_i + k; // Loop over all combinations of samples. We will maintain the iterator ranges so that // within the inner for loop we have: // [begin_i, end_i) == the range in edges that contains neighbors of samples[i] // [begin_j, end_j) == the range in edges that contains neighbors of samples[j] for (unsigned long i = 0; i+1 < samples.size(); ++i) { begin_j = begin_i; end_j = end_i; for (unsigned long j = i+1; j < samples.size(); ++j) { begin_j += k; end_j += k; const float dist = dist_funct(samples[i], samples[j]); if (dist < worst_dists[i]) { *iterator_of_worst(begin_i, end_i) = sample_pair(i, j, dist); worst_dists[i] = iterator_of_worst(begin_i, end_i)->distance(); } if (dist < worst_dists[j]) { *iterator_of_worst(begin_j, end_j) = sample_pair(i, j, dist); worst_dists[j] = iterator_of_worst(begin_j, end_j)->distance(); } } begin_i += k; end_i += k; } // sort the edges so that duplicate edges will be adjacent std::sort(edges.begin(), edges.end(), &order_by_index); // if the first edge is valid if (edges[0].index1() < samples.size()) { // now put edges into out while avoiding duplicates and any remaining invalid edges. out.reserve(edges.size()); out.push_back(edges[0]); for (unsigned long i = 1; i < edges.size(); ++i) { // if we hit an invalid edge then we can stop if (edges[i].index1() >= samples.size()) break; // if this isn't a duplicate edge if (edges[i] != edges[i-1]) { out.push_back(edges[i]); } } } } // ---------------------------------------------------------------------------------------- template < typename vector_type > bool contains_duplicate_pairs ( const vector_type& pairs ) { vector_type temp(pairs); std::sort(temp.begin(), temp.end(), &order_by_index); for (unsigned long i = 1; i < temp.size(); ++i) { // if we found a duplicate if (temp[i-1] == temp[i]) return true; } return false; } // ---------------------------------------------------------------------------------------- template < typename vector_type > unsigned long max_index_value_plus_one ( const vector_type& pairs ) { if (pairs.size() == 0) { return 0; } else { unsigned long max_idx = 0; for (unsigned long i = 0; i < pairs.size(); ++i) { if (pairs[i].index1() > max_idx) max_idx = pairs[i].index1(); if (pairs[i].index2() > max_idx) max_idx = pairs[i].index2(); } return max_idx + 1; } } // ---------------------------------------------------------------------------------------- template < typename vector_type > void remove_long_edges ( vector_type& pairs, float distance_threshold ) { vector_type temp; temp.reserve(pairs.size()); // add all the pairs shorter than the given threshold into temp for (unsigned long i = 0; i < pairs.size(); ++i) { if (pairs[i].distance() <= distance_threshold) temp.push_back(pairs[i]); } // move temp into the output vector temp.swap(pairs); } // ---------------------------------------------------------------------------------------- template < typename vector_type > void remove_short_edges ( vector_type& pairs, float distance_threshold ) { vector_type temp; temp.reserve(pairs.size()); // add all the pairs longer than the given threshold into temp for (unsigned long i = 0; i < pairs.size(); ++i) { if (pairs[i].distance() >= distance_threshold) temp.push_back(pairs[i]); } // move temp into the output vector temp.swap(pairs); } // ---------------------------------------------------------------------------------------- template < typename vector_type > void remove_percent_longest_edges ( vector_type& pairs, double percent ) { // make sure requires clause is not broken DLIB_ASSERT( 0 <= percent && percent < 1, "\t void remove_percent_longest_edges()" << "\n\t Invalid inputs were given to this function." << "\n\t percent: " << percent ); std::sort(pairs.begin(), pairs.end(), &order_by_distance); const unsigned long num = static_cast<unsigned long>((1.0-percent)*pairs.size()); // pick out the num shortest pairs vector_type temp(pairs.begin(), pairs.begin() + num); // move temp into the output vector temp.swap(pairs); } // ---------------------------------------------------------------------------------------- template < typename vector_type > void remove_percent_shortest_edges ( vector_type& pairs, double percent ) { // make sure requires clause is not broken DLIB_ASSERT( 0 <= percent && percent < 1, "\t void remove_percent_shortest_edges()" << "\n\t Invalid inputs were given to this function." << "\n\t percent: " << percent ); std::sort(pairs.rbegin(), pairs.rend(), &order_by_distance); const unsigned long num = static_cast<unsigned long>((1.0-percent)*pairs.size()); // pick out the num shortest pairs vector_type temp(pairs.begin(), pairs.begin() + num); // move temp into the output vector temp.swap(pairs); } // ---------------------------------------------------------------------------------------- } #endif // DLIB_GRAPH_CrEATION_H__