OpenCV 3.0 NormalBayesClassifier 预测错误

Question

我在 C++ 中使用 OpenCV 3.0.0 来训练正态贝叶斯分类器。我训练了分类器，并将其存储到一个文件中。现在我正在尝试使用分类器来预测响应。以下是相关代码：

Ptr<TrainData> trainingData = TrainData::loadFromCSV(trainingDataFile,0);
std::cout << "Training data has been loaded.\n";
Mat samples = trainingData->getTrainSamples();
Mat correctResponses = trainingData->getTrainResponses();

std::cout << "Loading NB classifier...\n";
Ptr<NormalBayesClassifier> nbClassifier = NormalBayesClassifier::load<NormalBayesClassifier>(trainingDataFile+"_trainedNBParams.dat");
std::cout << "NB classifier loaded.\n";

Mat receivedResponses;
std::cout << "Classifying data using NB classifier...\n";
nbClassifier->predict(samples,receivedResponses); //error here
std::cout << "Classification complete.\n";

我收到以下错误：

OpenCV Error: Null pointer (When the number of input samples is >1, the output vector of results must be passed) in predictProb, file /home/keyur/OpenCV/opencv-3.0.0/modules/ml/src/nbayes.cpp, line 318
terminate called after throwing an instance of 'cv::Exception'
  what():  /home/keyur/OpenCV/opencv-3.0.0/modules/ml/src/nbayes.cpp:318: error: (-27) When the number of input samples is >1, the output vector of results must be passed in function predictProb

我不明白应该如何将输出矩阵传递给函数。文档相当模糊，其他 StackOverflow 答案对这个特定问题没有帮助。

文档链接： http://docs.opencv.org/3.0.0/d4/d8e/classcv_1_1ml_1_1NormalBayesClassifier.html

Answer 1

该错误是由于 OpenCV 3.0.0 中的错误导致的，该错误已在 OpenCV 3.1.0 中修复。