Robert Sachunsky
|
25e967397d
|
exit early if no text regions found (to avoid segfault)
|
2 weeks ago |
|
Robert Sachunsky
|
21efea8711
|
no del on function argument
|
2 weeks ago |
|
Robert Sachunsky
|
5e0c1da711
|
simplify
|
2 weeks ago |
|
Robert Sachunsky
|
54cb15056b
|
do_image_rotation / return_deskew_slop: avoid code duplication, simplify via mp.Pool
|
2 weeks ago |
|
Robert Sachunsky
|
6fe02df973
|
do_image_rotation: fix f93fa12 (do return results)
|
2 weeks ago |
|
Robert Sachunsky
|
d68017037c
|
do_prediction: trigger GC to avoid CUDA OOM
|
2 weeks ago |
|
Robert Sachunsky
|
ad748d0039
|
do_prediction: avoid code duplication
|
2 weeks ago |
|
Robert Sachunsky
|
c3163caefd
|
avoid indentation
|
3 weeks ago |
|
Robert Sachunsky
|
055463d23a
|
avoid indentation
|
3 weeks ago |
|
Robert Sachunsky
|
aaea2ef463
|
simplify
|
3 weeks ago |
|
Robert Sachunsky
|
3d88b207fc
|
run: log instead of print
|
3 weeks ago |
|
Robert Sachunsky
|
a520bd1f77
|
wrap extremely long lines
|
3 weeks ago |
|
Robert Sachunsky
|
cd4e426977
|
avoid indentation (skip_layout_and_reading_order)
|
3 weeks ago |
|
Robert Sachunsky
|
5b82320707
|
avoid indentation
|
3 weeks ago |
|
Robert Sachunsky
|
9f12fa241d
|
log-level: only set 'eynollah' logger level
|
3 weeks ago |
|
Robert Sachunsky
|
14beb46224
|
simplify loading models w/o dir_in mode
|
3 weeks ago |
|
Robert Sachunsky
|
329fac23f6
|
do not reload enhancement model in dir_in mode, simplify
|
3 weeks ago |
|
Robert Sachunsky
|
3b9a29bc5c
|
simplify dir_in conditionals
|
3 weeks ago |
|
Robert Sachunsky
|
7ae64f3717
|
RO model: do not reload when in dir_in mode
|
3 weeks ago |
|
Robert Sachunsky
|
f765e2603b
|
move Torch to optional dependencies (to avoid clash with TF over CuDNN)
|
3 weeks ago |
vahidrezanezhad
|
871d7bfc5a
|
fixed: machine based reading order cause tuple index out of range error if number of textregion is one.
|
3 weeks ago |
|
vahidrezanezhad
|
6aad006f4c
|
filter textregions without textline
|
3 weeks ago |
kba
|
1083d1c7fb
|
gha: try to free disk space
|
4 weeks ago |
|
vahidrezanezhad
|
8014a9e416
|
Update Makefile
|
1 month ago |
|
vahidrezanezhad
|
3000255a24
|
Update Makefile
|
1 month ago |
|
vahidrezanezhad
|
1746920275
|
Update Makefile
|
1 month ago |
|
vahidrezanezhad
|
b622494f34
|
new table detection model is integrated
|
1 month ago |
|
vahidrezanezhad
|
d9f79c3404
|
fixing IndexError by reading order detection
|
1 month ago |
|
vahidrezanezhad
|
5fa8ca46a4
|
updating requirements
|
1 month ago |
|
vahidrezanezhad
|
ce5b611296
|
tests are passed - new models by the way should be uploaded
|
1 month ago |
|
vahidrezanezhad
|
f43c49c508
|
textlines of drop capitals are connected to corresponding textline if possible otherwise they are inserted in corresponding textregion
|
1 month ago |
|
vahidrezanezhad
|
22b0b07a73
|
drop capital and marginals extraction is updated
|
1 month ago |
Clemens Neudecker
|
1ae77e61c8
|
Update requirements.txt
|
1 month ago |
|
vahidrezanezhad
|
8409de0e58
|
sbb_binarization is integrated into eynollah works in framework of ocrd - sbb_binarization in ocrd works for individual images by the way as standalone flowing from directory can be used now. For eynollah in ocrd framework I have added -light version as default parameter.
|
1 month ago |
|
vahidrezanezhad
|
0914b5ff8a
|
resolve merge conflict of main branch with machine based reading order branch
|
2 months ago |
|
vahidrezanezhad
|
6aee70d0cd
|
Resolve merge conflict of main and machine based reading order branch
|
2 months ago |
|
vahidrezanezhad
|
bceeeb56c1
|
Merge pull request #138 from qurator-spk/extracting_images_only
Extracting images only
|
2 months ago |
|
vahidrezanezhad
|
f7e5fb917f
|
resolving merge conflict of machine based reading order and extracting only images branches
|
2 months ago |
|
vahidrezanezhad
|
751b0102f7
|
updating early layout inference for light version
|
2 months ago |
|
vahidrezanezhad
|
e796a99c5c
|
updating inference for early layout in the case of documents with number of columns bigger than 2
|
2 months ago |
|
vahidrezanezhad
|
438df52287
|
updating
|
2 months ago |
|
vahidrezanezhad
|
90ee2d61dc
|
textline segmentation is masked with drop capitals
|
2 months ago |
|
vahidrezanezhad
|
5037e9896d
|
Merge branch 'machine_based_reading_order_integration' of https://github.com/qurator-spk/eynollah into machine_based_reading_order_integration
|
2 months ago |
|
vahidrezanezhad
|
82281bd6cf
|
fixing a bug occuring with reading order + Slro option with no patch textline model and thresholding artificial class
|
2 months ago |
|
vahidrezanezhad
|
328d33e3dc
|
Temporary commit – textline prediction without patches
|
2 months ago |
|
vahidrezanezhad
|
70772d4104
|
binarization as a standalone command
|
2 months ago |
|
vahidrezanezhad
|
f93fa12441
|
doing more multiprocessing in order to make the process faster
|
2 months ago |
|
vahidrezanezhad
|
3ef4eac24c
|
textlines of textregions are extracted in a faster way + early layout for all documents is done with no patches model and on rgb input
|
2 months ago |
|
vahidrezanezhad
|
1da4b7f589
|
updating light version
|
3 months ago |
|
vahidrezanezhad
|
543ed4bc38
|
-light version need -tll to be enabled otherwise the process will be ended.
|
3 months ago |