How to use labeled faces in the wild data set for face recognition algorithms?

Hello I am new to face recognition techniques using matlab and that also using neural networks .

I have downloaded the entire zip folder of the original images of the label faces in the wild dataset.

It has extracted itself into various subfolders in which each folder contains a particular individual's name . Now kindly tell me how to use it to extract features from each and every image to be used as a training set for my neural network face recognition algorithm .

Suppose I want to extract the SIFT features of each and every image so how do I do that by looping in to the folders to get the features of each and every image and also up to how to divide the entire dataset into training and testing phase for my neural network.

There are some individuals who have only one image so I cannot use the same image of the person during the testing phase.how to overcome the is problem.

Kindly give me a brief idea on how to proceed.

[Greg Heath]

If there is only one image you can synthesize as many more as you want by adding random noise.

Answer 1

[zhuang linqiao]

i have the same question with you and how to solve it?