The calculating of CBR
xuanloc088 opened this issue · comments
First, I would like to thank you for providing the code so that we can understand your paper more clearly. However, I have one concern about this code is that the way you calculate the CBR. In the code, CBR= feature.numel() / 2 / input_image.numel(). I understand why you take the feature / input image, but I don't understand why you take the feature.numel()/2. If you don't mind, can you tell me the reason why, that would be great. Once again, thank you for the great work.
Because our paper uses complex signals for transmission.
Thank you for your answer, I have looked into the channel code and realize that you divide the first half of the feature to real components and the other half into imaginary components. I understand it now. Thank you very much