About Booksorber
Today I've made some sample pictures to use with booksorber, and while it looks promising the application still needs loads of work.
The solution for setting up a camera on a tripod, and using this applications is regarded with mixed feelings:
It's javabased and slow.
The boundary detection (edges of the book) isn't working (yet)
while you manually need to set the boundary, the pages won't become straight unless you have loads of patience, did I mention it's slow? Moving a node feels like pulling a 100kg ball, ok perhaps only 50kg.
Some text in one page became black, while some other became light gray.
The automatic finger removal doesn't work.
I had to manually fix all the pages (I took 4 pictures) after it's processing job.
It takes JPG as input, and throws out JPG and PDF. Tbh I'd prefered the output to be TIF or PNG.
While all automatic, it doesn't do white balancing on images, so images are in the hands of the light when taking the picture.
[Some good things]
The 2 pages that came out decent were quite good
it generates a PDF when done.
It runs on any computer with Java installed
the verdict of booksorber version 1.1.3 - it doesn't live up to it's expectations of what they say on their website
The job could be made better and faster using a scanner and image processing software (did I mention it was slow?)
Hardware used:
Canon 700D with flash
Canon wireless remote trigger
Tripod
Black cardboard
The demo version makes water marks.. but even though you can test it yourself :-)
