The input should be one image that consists of several similar images. Maybe have and arrangement like nx1 or 1xn. ImageDiffer can help you to figure out the main difference between them.
usage: imageDiffer [-h] -f IMAGE_PATH [-d {h,v}] [-c COUNT]
[-b BACKGROUND_COLOR]
Process image diff
optional arguments:
-h, --help show this help message and exit
-f IMAGE_PATH, --file IMAGE_PATH
the input image path
-d {h,v}, --direction {h,v}
h: stands for horizontal; v: stands for vertical;
default to h
-c COUNT, --count COUNT
the count to split the image into; default to 2
-b BACKGROUND_COLOR, --bgcolor BACKGROUND_COLOR
the r,g,b color of background; default to 255,255,255
Here's the example to use the default configuration to find difference.
python imageDiffer.py -f input.jpg