Do you know if there is a file size limit ?

[aurelien-slodzian]

Hello,

Thank you for sharing your work; very useful !

I have a set of very large TIF images; not large in the sense of KB but in pixels: about 18Kx30K pixels (scans of A0 documents).

When I use ocrmac with these files, it provides very few results; if however I cut the image in pieces I get a whole lot more text from the image.

I guess it might be a limitation of the underlying library but found no information about such limitations.

Any hint ?

Regards,
Aurélien

[straussmaximilian]

Hi,
I am not aware of any limitations w. regards of size. Can you share some test files to investigate?

Tiling makes sense but then of course there is the issue of stitching together again..

[aurelien-slodzian]

PLAN_2024-10-31_10-54-38-rotated270.tif.zip

This is an example; if I try with the whole image, I get very few results from the OCR (both vision kit and live text); but if I cut it in e.g. 9 tiles, then I get excellent results... (reconciling the coordinates is rather easy, but if I expect to retrieve all text then the tiles should in fact overlap, which renders things a bit messier to stitch together)