Closed sacdallago closed 8 years ago
I wrote a code for this issue, it's in Jupyter, the file is called Normalization.ipynb. The code is taking a bunch of files from the folder Normalization_Try and is scaling up to a value I set up in the beginning (Now is set to 300, we can decide later which one we want to use, or if we are taking the median or mean). The new files are saved in the folder Normalization_Try_Resized.
This should also be done
Write python script to crop
train_pictures
into squares of X*X .Anyone has a suggestion for the size X we should use? I think it would make sense to have the python script select the size automatically by getting all widths and all heights and choosing: