Java OpenCV培训SVM错误-输入样本必须为一维向量

问题描述

我想在给定的图像上检测笑容，但出现错误的参数错误:

I want to detect smile on given image, but I get a bad argument error:

OpenCV Error: Incorrect size of input array (Input sample must be 1-dimensional vector) in cvPreparePredictData, file ..\..\..\..\opencv\modules\ml\src\inner_functions.cpp, line 1107
Exception in thread "main" CvException [org.opencv.core.CvException: cv::Exception:    ..\..\..\..\opencv\modules\ml\src\inner_functions.cpp:1107: error: (-201) Input sample must be 1-dimensional vector in function cvPreparePredictData
]

在detectSmile(Mat face)方法的最后一行.整个代码:

at the last line of detectSmile(Mat face) method. The whole code:

public class SmileDetector {
private CascadeClassifier faceDetector;
private Mat image;
private Mat classes;
private Mat trainingData;
private Mat trainingImages;
private Mat trainingLabels;
CvSVM clasificador;


public void detectSmile(String filename) {
    init();
    detectFace(filename);
    Mat face = Highgui.imread("output.png", Highgui.CV_LOAD_IMAGE_GRAYSCALE);
    detectSmile(face);
}

private void init() {
    System.loadLibrary(Core.NATIVE_LIBRARY_NAME);
    faceDetector = new CascadeClassifier(new File(
            "src/main/resources/haarcascade_frontalface_alt.xml").getAbsolutePath());
    classes = new Mat();
    trainingData = new Mat();
    trainingImages = new Mat();
    trainingLabels = new Mat();
}

private void detectFace(String filename) {
    image = Highgui.imread(filename);
    MatOfRect faceDetections = new MatOfRect();
    faceDetector.detectMultiScale(image, faceDetections);
    System.out.println(String.format("Detected %s faces", faceDetections.toArray().length));
    for (Rect rect : faceDetections.toArray()) {
        Core.rectangle(image, new Point(rect.x, rect.y), new Point(rect.x + rect.width, rect.y + rect.height),
                new Scalar(0, 255, 0));
    }
    Highgui.imwrite("output.png", image.submat(faceDetections.toArray()[0]));
}

private void trainSVM() {
    trainPositive();
    trainNegative();
    trainingImages.copyTo(trainingData);
    trainingData.convertTo(trainingData, CvType.CV_32FC1);
    trainingLabels.copyTo(classes);
}

private void trainPositive() {
    Mat img = new Mat();
    Mat con = Highgui.imread("D:\\cybuch\\workspace\\facerecognizer\\src\\main\\resources\\happy.png", Highgui.CV_LOAD_IMAGE_GRAYSCALE);
    con.convertTo(img, CvType.CV_32FC1,1.0/255.0);
    trainingImages.push_back(img.reshape(1, 1));
    trainingLabels.push_back(Mat.ones(new Size(1, 1), CvType.CV_32FC1));
}

private void trainNegative() {
    Mat img = new Mat();
    Mat con = Highgui.imread("D:\\cybuch\\workspace\\facerecognizer\\src\\main\\resources\\sad.png", Highgui.CV_LOAD_IMAGE_GRAYSCALE);
    con.convertTo(img, CvType.CV_32FC1,1.0/255.0);
    trainingImages.push_back(img.reshape(1,1));
    trainingLabels.push_back(Mat.zeros(new Size(1, 1), CvType.CV_32FC1));
}

private void detectSmile(Mat face) {
    trainSVM();
    CvSVMParams params = new CvSVMParams();
    params.set_kernel_type(CvSVM.LINEAR);
    clasificador = new CvSVM(trainingData, classes, new Mat(), new Mat(), params);
    clasificador.save("svm.xml");
    clasificador.load("svm.xml");
    System.out.println(clasificador);
    Mat out = new Mat();
    face.convertTo(out, CvType.CV_32FC1);
    out.reshape(1, 1);
    System.out.println(out);
    System.out.println(clasificador.predict(out));
}
}

我仅使用一个训练图像样本，但稍后将对其进行重写.目前，我只想进行预测.

I use only one sample of training images, but I'll rewrite it later. For now I'd like just to make predictions working.

推荐答案

重塑 返回另一个Mat，它不能就地 工作.

reshape returns another Mat, it does not work in-place.

因此，将您的代码更改为:

so, change your code to :

out = out.reshape(1, 1);

(或clasificador.predict(out.reshape(1,1)))

( or clasificador.predict(out.reshape(1,1)) )

(不用担心，如果我每个人都有一分钱，走进那个坑，我就会很有钱)

(no fear, if i had a penny for everyone, who stepped into that pit, i'd be rich)

这篇关于Java OpenCV培训SVM错误-输入样本必须为一维向量的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！