Hi Clement, the purpose of the app is to screen for things you don't know are there (unexpected duplicate images). If you think the image is edited, the app cannot be used to confirm or deny that. It will work best on biomedical literature.
I agree with your comment, the image looks edited. No apps like this should be used to confirm an accusation. They are tools, like a magnifying glass, it can help you see things that are hard to spot, it cannot tell you anything about what you see.
Most papers do not have duplicate images. Not all duplicate or overlapping images will be detected by this app. It is best used as a supportive tool for investigating papers that you are interested in. If it helps you find something, great!
Very good! Then you have already seen it, what do you need the tool for!? All methods must balance sensitivity and specificity. Without knowing exactly what paper you are scanning, I can't say more.
Comments
I tried using this paper but it doesn't find anomalies :
https://pubpeer.com/publications/9535FEB471C40A2FB8655C52B4AAF5
Author answered (through editors) that's an artefact of the interpolation process
I'm quasi-sure that no and strongly suspect the image was edited
Hope your tool will help then !