Skip to content

Recognize multiple rectangles on a page without clearing results inbetween #4366

@bleze

Description

@bleze

Your Feature Request

Using SetRectangle and Recognize in a loop will clear what has been found between each run resulting in only the last result being returned. It would be nice if it would nice it was possible to keep all results by telling it not to clear. Alternatively give it a range of rectangles and not just one. The latter is probably faster, since overhead can be reduced.

In turn when inserting result via AddImage in a PDFRenderer, the page is automatically advanced to the next, so only one result can be inserted per page. It would also be nice if at least it was possible to tell it not/or when to advance, or which page you are adding to, thereby build the pages from OCR results.

These features would help me achieve what I estimate is 80% of the time spent, or put it in another way 5x speed. Unsure about memory saved, but might help here also.

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions