Use the image_processor associated with the pre-trained model to obtain the following information:

Image mean and standard deviation with which the video frame pixels will be normalized.