Image Processing Help - Find differences in scanned document images

with python and opencv
We are creating a program that analyzes the difference between the first scanned image of a paper document and the second scanned image, and highlights where the text is written in red.

Due to the physical curvature during the first scan and the second scan, even if it is slightly distorted, all letters are recognized as differences, and it is difficult.

Is it possible?