Difference from raw image and image in npy

Question

I am working with some EDF (European Data Format) images, and I have the following problem: if I load the files in a npy array, and I compare a certain array element with the corresponding raw file, I get that

• The files look the same BUT
• The difference is not 0. Plotting Image_from_stack - Ram_image, I get a striped value distribution (see image). Does anyone have a suggestion on what could be the cause for this, and how to fix it?

To make things more interesting, the difference changes from image to image, but it always shows a striped pattern.

I am working in python.

Answer 1

A note for future readers: the problem explained above was related to a scientific programming script running on a high performance computing machine. The script was using a substantial amount of memory (up to 100 GB).

My guess is that the striped pattern effect presented above is related to such anomalous memory requirements. After rebooting the machine I couldn't replicate the problem.

So in case you see something similar, check the memory usage. If it's very high, give reboot a chance!