问题描述
我正在尝试使用opencv执行面部识别.为此，训练图像和测试图像必须具有相同的大小. 所以我用了这个功能:
I am trying to perform facial recogntion using opencv. For this the training and test images have to be of the same size. So I used this function:
resize(testing,output, images[0].size(),0,0,INTER_NEAREST);
其中测试是来自网络摄像头的Mat，输出是调整大小的图像，images [0] .size()是我的矢量图像数据库. 但是我收到以下错误:
Where testing is my Mat from the webcam , Output is the resized image , images[0].size() is my database of vector images. However I get the following error:
OpenCV Error: Bad argument (Wrong shapes for given matrices. Was size(src) = (1,921600),
size(W) = (307200,6).) in subspaceProject,
file /home/someplae/OpenCV-2.4.2/modules/contrib/src/lda.cpp,
line 187
terminate called after throwing an instance of 'cv::Exception'
what(): /home/someplace/OpenCV-2.4.2/modules/contrib/src/lda.cpp:187:
error: (-5) Wrong shapes for given matrices. Was size(src) = (1,921600), size(W) = (307200,6).
in function subspaceProject
推荐答案
尽管您没有提供足够的代码来进行验证，但是我可以从给定的元素数量中推断出该问题.您的测试矩阵是921600 = 3 * 307200 ...所以我想您的输入矩阵不是灰度矩阵，但仍然是BGR表示形式.
Although you don't provide enough code to verify, I can infer the problem from the given number of elements. Your test matrix is 921600 = 3 * 307200... So I guess your input matrix is not a grayscale matrix, but still is in BGR representation.
您应该使用 cv :: cvtColor 将矩阵转换为灰度，然后再调整其大小:
You should use cv::cvtColor to convert the matrix to grayscale, before resizing it:
这篇关于OPENCV中给定矩阵的形状错误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！