| Age | Commit message (Collapse) | Author |
|---|
|
|
|
|
|
|
|
|
|
If content is very light or sparse it may be better to not wipe at all
than wipe almost all of the content leaving a small strip. This is done
now by aborting the wipe if the detected content takes up less than the
minimum % of the page (default is 30%).
|
|
|
|
rather than average line confidence
|
|
rather than float, for easier comparison
|
|
|
|
|
|
|
|
|
|
Surprisingly opening an image takes a significant amount of the
total processing time, so this actually saves quite a bit of time
in the grand scheme of things.
|
|
cmds as they were defined in global package space
|
|
|
|
|
|
Also unify everything else under preproc/
Note that the UsefulImg interface should be used by the main
functions, to simplify things, but this hasn't been done yet.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This uses integral image calculations, so they have been exported in the binarization package
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Note that there are some very small differences to the output compared
to the basic algorithm, but this doesn't make much difference.
This is due to minor differences with the standard deviation
calculation throughout, and with mean calculation at edges, for reasons
I'm unclear about.
WIP integral image speedup. mean is working
Very WIP, but mean is perfect once full window is used
Integral version all working!
Remove debugging info
Organise code better
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
processed
|
