OpenCV3.3中 K-最近邻法(KNN)接口简介及使用

OpenCV 3.3中给出了K-最近邻(KNN)算法的实现，即cv::ml::Knearest类，此类的声明在include/opecv2/ml.hpp文件中，实现在modules/ml/src/knearest.cpp文件中。其中：

(1)、cv::ml::Knearest类：继承自cv::ml::StateModel，而cv::ml::StateModel又继承自cv::Algorithm;

(2)、create函数：为static，new一个KNearestImpl用来创建一个KNearest对象；

(3)、setDefaultK/getDefaultK函数：在预测时，设置/获取的K值；

(4)、setIsClassifier/getIsClassifier函数：设置/获取应用KNN是进行分类还是回归；

(5)、setEmax/getEmax函数：在使用KDTree算法时，设置/获取Emax参数值；

(6)、setAlgorithmType/getAlgorithmType函数：设置/获取KNN算法类型，目前支持两种：brute_force和KDTree；

(7)、findNearest函数：根据输入预测分类/回归结果。

关于KNN算法介绍可以参考： http://blog.csdn.net/fengbingchun/article/details/78464169  

以下是从数据集MNIST中提取的40幅图像，0,1,2,3四类各20张，每类的前10幅来自于训练样本，用于训练，后10幅来自测试样本，用于测试，如下图：

关于MNIST的介绍可以参考：  http://blog.csdn.net/fengbingchun/article/details/49611549 

测试代码如下：

#include "opencv.hpp"
#include 
#include 
#include 
#include 
#include 
#include 
#include "common.hpp"

/////////////////////////////////////////// K-Nearest Neighbor(KNN) //////////////////////////////////////
int test_opencv_knn_predict()
{
	const int K{ 3 };
	cv::Ptr knn = cv::ml::KNearest::create();
	knn->setDefaultK(K);
	knn->setIsClassifier(true);
	knn->setAlgorithmType(cv::ml::KNearest::BRUTE_FORCE);

	const std::string image_path{"E:/GitCode/NN_Test/data/images/digit/handwriting_0_and_1/"};
	cv::Mat tmp = cv::imread(image_path + "0_1.jpg", 0);
	const int train_samples_number{ 40 }, predict_samples_number{ 40 };
	const int every_class_number{ 10 };

	cv::Mat train_data(train_samples_number, tmp.rows * tmp.cols, CV_32FC1);
	cv::Mat train_labels(train_samples_number, 1, CV_32FC1);
	float* p = (float*)train_labels.data;
	for (int i = 0; i < 4; ++i) {
		std::for_each(p + i * every_class_number, p + (i + 1)*every_class_number, [i](float& v){v = (float)i; });
	}

	// train data
	for (int i = 0; i < 4; ++i) {
		static const std::vector digit{ "0_", "1_", "2_", "3_" };
		static const std::string suffix{ ".jpg" };

		for (int j = 1; j <= every_class_number; ++j) {
			std::string image_name = image_path + digit[i] + std::to_string(j) + suffix;
			cv::Mat image = cv::imread(image_name, 0);
			CHECK(!image.empty() && image.isContinuous());
			image.convertTo(image, CV_32FC1);

			image = image.reshape(0, 1);
			tmp = train_data.rowRange(i * every_class_number + j - 1, i * every_class_number + j);
			image.copyTo(tmp);
		}
	}

	knn->train(train_data, cv::ml::ROW_SAMPLE, train_labels);

	// predict datta
	cv::Mat predict_data(predict_samples_number, tmp.rows * tmp.cols, CV_32FC1);
	for (int i = 0; i < 4; ++i) {
		static const std::vector digit{ "0_", "1_", "2_", "3_" };
		static const std::string suffix{ ".jpg" };

		for (int j = 11; j <= every_class_number+10; ++j) {
			std::string image_name = image_path + digit[i] + std::to_string(j) + suffix;
			cv::Mat image = cv::imread(image_name, 0);
			CHECK(!image.empty() && image.isContinuous());
			image.convertTo(image, CV_32FC1);

			image = image.reshape(0, 1);
			tmp = predict_data.rowRange(i * every_class_number + j - 10 - 1, i * every_class_number + j - 10);
			image.copyTo(tmp);
		}
	}

	cv::Mat result;
	knn->findNearest(predict_data, K, result);
	CHECK(result.rows == predict_samples_number);

	cv::Mat predict_labels(predict_samples_number, 1, CV_32FC1);
	p = (float*)predict_labels.data;
	for (int i = 0; i < 4; ++i) {
		std::for_each(p + i * every_class_number, p + (i + 1)*every_class_number, [i](float& v){v = (float)i; });
	}

	int count{ 0 };
	for (int i = 0; i < predict_samples_number; ++i) {
		float value1 = ((float*)predict_labels.data)[i];
		float value2 = ((float*)result.data)[i];
		fprintf(stdout, "expected value: %f, actual value: %f\n", value1, value2);

		if (int(value1) == int(value2)) ++count;
	}
	fprintf(stdout, "when K = %d, accuracy: %f\n", K, count * 1.f / predict_samples_number);

	return 0;
}

测试结果如下：

GitHub： https://github.com/fengbingchun/NN_Test