Thank you very much for your work. My question is whether there is data available about the windowing settings used to converting CT scans (typically in Hounsfield Units) to images with intensity levels given as integers in the range 0 to 255. When searching for lung nodules a window with center at -200 HU and width 2000 HU may be appropriate (that is what I have used in the past), but if we want to use a pretrained model such as ResNet50 trained on RadImageNet it makes sense to feed it with images obtained with the same windowing settings of the images in the training set used for pretraining the model. What window settings were used for obtaining the RadImageNet images of lung nodules?
Thank you in advance.
Thank you very much for your work. My question is whether there is data available about the windowing settings used to converting CT scans (typically in Hounsfield Units) to images with intensity levels given as integers in the range 0 to 255. When searching for lung nodules a window with center at -200 HU and width 2000 HU may be appropriate (that is what I have used in the past), but if we want to use a pretrained model such as ResNet50 trained on RadImageNet it makes sense to feed it with images obtained with the same windowing settings of the images in the training set used for pretraining the model. What window settings were used for obtaining the RadImageNet images of lung nodules?
Thank you in advance.