Opened 20 months ago
Last modified 20 months ago
#10378 new enhancement
Feature Request: positional ocr of dvd and bd subs (in combination with tesseract)
Reported by: | techguru | Owned by: | |
---|---|---|---|
Priority: | important | Component: | undetermined |
Version: | unspecified | Keywords: | frame data ocr |
Cc: | Blocked By: | ||
Blocking: | Reproduced by developer: | no | |
Analyzed by developer: | no |
Description
if you combine ffmpeg and tesseract(and maybe opencv)....it should be possible to get positional frame data of the text in the frame
and its been a BADLY needed feature for those of us that do OCRing of dvd and bd subs for many many years now
I have attached a very good example of the sub files that has a lot of this positional use
(and yes I understand theres no positional data in the sub file...thats where ffmpeg in conjunction with tesseract should come in...it should be able to find that data in relation to the size of the frame)
the output subtitle would have to be .ass since its one of the more popular types with positional awareness
Attachments (1)
Change History (4)
by , 20 months ago
comment:2 by , 20 months ago
no its not...it can only ocr it and put it at the bottom
it doesn't know positioning
comment:3 by , 20 months ago
NISKE...author of SE..WANTS to add this feature...but the frame needs to be processed by a ffmpeg/libav to pass that data to tesseract first
sub files