Robert Sachunsky

> instead of joining them (expanding the textregions)? I'll do a little demo PR for that today...

The problem is still there with #173:  Changing the prior text mask dilation https://github.com/qurator-spk/eynollah/blob/a2359ea4c45e6eb3eb704d79b1b011889c3ec8db/src/eynollah/eynollah.py#L1893 from kernel size 2 → 5 results in this:  (i.e. more of the original...

Ok, I see what you mean. (The marginals get merged with the column via the text lines.) But couldn't we simply do a (foreground-) majority vote on each text line,...

Getting back to your example: with the recent changes in (the marginalia detection of) 0.5.0, that page actually does not work correctly anymore: So the marginalia in the middle now...

Sorry for the force-push! I had accidentally rebased back to https://github.com/qurator-spk/eynollah/pull/206/commits/a2a06a8828e50f5437d4aba773361b7137cc97c7 (which now became https://github.com/qurator-spk/eynollah/pull/206/commits/cd35241e816acc7e2083dc31d99f376a8877904b). Still have not addressed the big TODO (which is coming shortly), but found some more...

Done! Let me explain… - [1a76ce1](https://github.com/qurator-spk/eynollah/pull/206/commits/1a76ce177dba69aa711b74e6c69022e4a5ebf27f) is needed because sometimes contours centers in float were exactly "between" bboxes in integer, e.g. cy=400.3 where one box ends on 400 and the...

Examples: | *original* | *final RO boxes* | *final RO* | | --- | --- | --- | | | | | | | | | |  | |...

@vahidrezanezhad thanks to your regression test, I was able to address remaining issues with another series of commits: - ee59a68 fixes a severe typo I made in 0.6 – me...

@vahidrezanezhad the new recursive RO algorithm is described briefly [here](https://github.com/bertsky/eynollah/blob/9fdae72e9620bd0ebd3bcef6bd8189fe8a003734/src/eynollah/utils/__init__.py#L1819-L1831), and its implementation [follows](https://github.com/bertsky/eynollah/blob/9fdae72e9620bd0ebd3bcef6bd8189fe8a003734/src/eynollah/utils/__init__.py#L1832-L1899) in just a few lines. Regarding the question of whether we should (as a general rule)...

#### So what does `processor.resolve_resource(name)` actually do? Well, … https://github.com/OCR-D/core/blob/47b77aa4249a93ebf9b279f43bb841c192facc20/src/ocrd/processor/base.py#L937-L938 … thus it uses `ocrd_utils.list_resource_candidates()` with the processor's `self.moduledir`. The latter (being installation-dependent) is only known to the processor's codebase,...