No self-introduction
Loading
Profile
Follow
codes (3)
Opencv test method for face recognition accuracy o
4.0
Application background Vs2010+opencv2.4.8, this is to measure the accuracy of face recognition code, the method of specific face recognition in the other code I uploaded. There are also two CPP, testLBPHaccuracy.cpp measured accuracy of LBP method, testLBPloadAccuracy.cpp is more streamlined, measured is direct download classifier is used to test the accuracy, the theory that two should be the same. Key Technology You can modify the CreateLBPHFaceRecognizer () parameter such as default parameters for this function
(1,8,4,4,DBL_MAX) and modify the training picture count (that is, modify the at.txt file) separately measured their accuracy and timing. With me changes the accuracy and time (including vs run time and runs on the Board on time) we can refer to, from memory, time, and accuracy rate to select the most suitable parameters.
tuomoding 7 0
1
Face recognition matlab absolutely available, not
4.0
Application background Face recognition matlab implementation, the use of ORL face database, there are 40 people, one person 10 picture, a total of 400. For the simple calculation, I only took the image of the first 10 people, in fact, the same. Code is placed inside a word document, which will only be placed in the current folder and then copy the document to the MATLAB. Key Technology Using the PCA algorithm:PCA based face recognition algorithm. Firstly, the application of the K-L transform, calculate value of the characteristics of the training face space, to choose eigenvalues, then constitute a new low dimensional orthogonal basis space, we will all face projection in the low dimensional space is then calculated with the recent facial images the face in the image to be measured, the final completion of face recognition.
tuomoding 0 0
1
Face recognition opencv, absolutely available
4.0
Application background Vs2010+opencv2.4.8, face recognition, ORL face database, there are two code, LBP.cpp using the LBP method to train the image to get the classifier (XML) and then test the test sample. LBPload.cpp can directly use LBP.CPP to generate a good classifier test to measure the picture, much faster than LBP.cpp. Key Technology Using the LBP algorithm, but also can be replaced by PCA and fish algorithm, but LBP faster and more memory is also a small number, there are test time. In order to facilitate, I directly determine the measured is thirteenth people, can be modified. Inside the data folder inside the at.txt and to be measured the picture path remember to modify the situation according to your situation.
tuomoding 0 0
1
No more~
Follows0
Fans0