> This would depends on the false positive / false negative rate. I'm not a doctor, but I am a physicist and former pro-photographer, what is noise and what is

> This would depends on the false positive / false negative rate. I'm not a doctor, but I am a physicist and former pro-photographer, what is noise and what is signal in an experiment whose output is an 'image' has nothing to do with what makes a photo look good to human eyes. Often the whole point of methods of visualization is to make the image look objectively bad so you can easily pick out the areas of interest by the fact they are an eyesore. Applying upscaling to those images will actively destroy vital data.