I would like to know if converting an image to gray scale is necessary step for all image pre processing techniques. I am using a neural network for face recognition. Is it really necessary for converting it into a gray scale or can I give color images also as input to neural networks?