Working of model in low-lit conditions

[AjayBhat29]

Hello @natanielruiz
I have a few doubts that I would like to clarify.

1. Is this model suitable for estimating head pose during low-lit and night vision conditions?
   If yes, is it expected to give similar accuracy as normal well-lit conditions?
   Also, are we supposed to include some changes for it to work in dim-lit conditions? If yes, could you please mention them?

2. In the code file the following snippet is used:

transformations = transforms.Compose([transforms.Scale(224), transforms.CenterCrop(224), transforms.ToTensor(), transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225])])

Is there any particular need to transform the image into 224 X 224?
If yes, could you please explain the need to do the same?

P.S. I am really grateful this implementation exists.