-
-
Notifications
You must be signed in to change notification settings - Fork 108
Description
Describe your problem:
Especially with logographic writing systems, using a different PSM mode might be necessary. As an example, NormCap can't detect singular Chinese characters. Tesseract can, if the PSM mode is (for example) 5 or 6. Additionally, the ability to change PSM mode would result in far cleaner text output for specific cases, such as storefront text, text on receipts, specific webpage layouts, book paragraphs, etc.
Solution you'd like to see:
An option to specify Tesseract's PSM mode would be very useful. If the "parse text" feature is supposed to be trying out different PSM modes already, then at least for me, it isn't doing much. Allowing the user to set the specific PSM mode would likely be the simplest way to improve overall results, and offer more use cases.
Alternatives you considered:
NormCap could try multiple different PSM modes automatically, and use whichever one gives the most accurate result for that specific capture. However, this would likely increase processing time, and may cause other issues, hence why I think letting the user choose the PSM mode would likely be the best option.
Additional information or remarks:
No response