The current page detection algorithm only works if the paper is rectangular.
In many cases, one may have pages to scan that have a folded angle. This common if you are trying to scan the second page of a document that is stapled at the left top corner. You get something like
__________
/ |
| |
| |
| |
| |
------------
The folded angle is enough to make the paper detection fail miserably (possibly because the algorithm looks for a rectangle and here you have a pentagon).
Steps to reproduce
Take a document made of at least two pages. Staple it at the top left corner. Fold it to see the second page to the max of its extension, by placing the first page on the back. You get a folded corner. Try to acquire the page.
Expected behaviour
The software should be able to detect the page from the background, regardless of the folded angle.
Actual behaviour
The paper detection algorithm cannot detect the page from the background.
kofaysi
commented
3 years ago 
The current page detection algorithm only works if the paper is rectangular.
In many cases, one may have pages to scan that have a folded angle. This common if you are trying to scan the second page of a document that is stapled at the left top corner. You get something like
The folded angle is enough to make the paper detection fail miserably (possibly because the algorithm looks for a rectangle and here you have a pentagon).
Steps to reproduce
Take a document made of at least two pages. Staple it at the top left corner. Fold it to see the second page to the max of its extension, by placing the first page on the back. You get a folded corner. Try to acquire the page.
Expected behaviour
The software should be able to detect the page from the background, regardless of the folded angle.
Actual behaviour
The paper detection algorithm cannot detect the page from the background.