Robert Sachunsky
01312c6369
recognize: delegate to core functions
2022-02-11 07:50:58 +01:00
Robert Sachunsky
5f23c03cd9
recognize: remove checkpoint param in favour of checkpoint_dir alone
2022-02-10 17:41:51 +01:00
Robert Sachunsky
11615be6b2
ocrd-tool.json: add model content-type, remove glob variant
2022-02-10 17:38:37 +01:00
76b34c50cb
📦 v1.0.3
2021-07-21 11:56:37 +02:00
34013ddb02
📝 Reduce process() docstring again
2021-07-02 20:24:20 +02:00
Robert Sachunsky
4c6d6655e1
improve process() docstring
2021-06-26 13:50:20 +02:00
Robert Sachunsky
3bde7cb37f
init from constructor not process(), use conventional name setup()
2021-06-26 13:50:00 +02:00
da98713e73
📦 v1.0.2
2021-02-09 20:03:49 +01:00
0869386ec4
🐛 Fix word and glyph coordinates
...
Fixes GH-57.
2021-02-09 19:57:03 +01:00
4cf25b8119
🎨 Rename input_channels variable to network_input_channels
2021-02-09 18:20:46 +01:00
c0902cdef5
Merge branch 'master' into image-features
2021-02-09 18:17:23 +01:00
1bb72cbaf1
📦 v1.0.1
2021-01-28 15:45:03 +01:00
53c94fea95
Merge pull request #53 from OCR-D/resolve-resources
...
Resolve resources
2021-01-27 14:09:29 +01:00
Konstantin Baierer
03f5e44e62
define default for checkpoint_dir, but allow checkpoint still
2021-01-27 13:59:45 +01:00
a014bab5b6
Merge pull request #49 from OCR-D/fix-48
...
check for empty line image, ht @andbue, fix #48
2021-01-20 15:37:50 +01:00
e7fb432e35
Merge pull request #52 from OCR-D/checkpoint_dir
...
Checkpoint dir
2021-01-19 11:59:15 +01:00
Konstantin Baierer
00e43b1d1f
use Processor.resolve_files to handle on-demand download of models via registry
2020-12-28 11:49:13 +01:00
Konstantin Baierer
fdd30ebb89
also add tensorflow version to --version output
2020-12-22 16:34:27 +01:00
Konstantin Baierer
d6804bd9c3
fix typos
2020-12-22 16:29:40 +01:00
Konstantin Baierer
83adfcfd5a
implement "checkpoint_dir" parameter as a simpler alternative to "checkpoint"
2020-12-22 15:20:45 +01:00
Konstantin Baierer
fe973e58db
add version of calamari in --version output
2020-12-17 18:07:38 +01:00
Konstantin Baierer
df530877dc
check for empty line image, ht @andbue, fix #48
2020-12-17 17:18:50 +01:00
448a5b0dbc
📦 v1.0.0
2020-11-25 12:56:50 +01:00
8fcd331fbd
Merge branch 'feat/update-calamari1'
2020-11-25 12:10:59 +01:00
Konstantin Baierer
e4982aff37
getLogger per method
2020-09-24 10:25:47 +02:00
Konstantin Baierer
f746b73fd0
use make_file_id and assert_file_grp_cardinality
2020-08-06 15:24:53 +02:00
86410110bc
📦 v0.0.7
2020-08-06 12:39:45 +02:00
7da45a0ec1
Set pcGtsId
...
Newest OCR-D validation checks PAGE-XML pcGtsId against METS file/@ID.
Set the pcGtsId here correctly.
Fixes #40 .
2020-08-06 12:31:47 +02:00
93190fae3b
⚡ Recognize more than one line at a time (Fixes gh#20)
2020-07-22 16:03:10 +02:00
123ee61a8b
v0.0.6
2020-02-13 16:04:17 +01:00
62e5e0c295
🐛 ocrd-tool.json: Fix GitHub url by s/kba/OCR-D
2020-02-13 16:00:58 +01:00
0334a35870
🐛 Sort predictions in exactly the same way, also when building the text
2020-02-12 17:18:37 +01:00
0c9e1f13c7
🐛 Sort predictions in exactly the same way to make sure we are correctly removing spaces
2020-02-12 16:38:45 +01:00
d2c843aa3f
📦 v0.0.5
2020-02-12 13:33:38 +01:00
cd8f6a5fcb
🐛 Use line id for debug message
2020-02-12 13:32:10 +01:00
5b6d8b3f41
🐛 Build line text on our own
...
Calamari does whitespace post-processing on prediction.sentence, while
it does not do the same on prediction.positions. Do it on our own to
have consistency.
Fixes GH-37.
2020-02-12 12:25:25 +01:00
4508e3ec47
📦 v0.0.4
2020-02-05 17:55:51 +01:00
b802b4deaf
✨ Allow configuring a cut off confidence value for glyph alternatives
2020-02-05 13:29:44 +01:00
ef3fb44fb5
✨ Allow controlling of output hierarchy level, e.g. only line, not words+glyphs
2020-02-05 13:02:10 +01:00
6f4736f8e4
✨ Do word segmentation as expected by OCR-D PAGE specs
2020-02-03 19:10:16 +01:00
0f9c94e7dc
🐛 Start with TextEquiv index=1 to adhere to OCR-D PAGE conventions
...
https://ocr-d.github.io/page#multiple-textequivs
2020-02-03 17:40:45 +01:00
909632493b
🚧 Add future TODOs
2020-02-03 17:37:19 +01:00
3149e1d9e0
📝 unwanted()
2020-02-03 15:33:38 +01:00
91cca1e1b8
📝 Document why we are using Unicode text segmentation to produce word results
2020-02-03 15:33:11 +01:00
2650189910
🧹 Add whitespace
2020-02-03 14:03:54 +01:00
f75426060e
🧹 Remove debugging print
2020-02-03 14:03:28 +01:00
decaa7b69f
🎨 Use polygon_from_x0y0x1y1 to build word/glyph polygon
2020-02-03 14:03:01 +01:00
2ccfc7b195
🎨 Set vim textwidth
2020-02-03 12:22:22 +01:00
507bc1ce5e
✨ Include proper word + glyph segmentation
2020-02-03 12:22:01 +01:00
24532f693a
🚧 Use character positions as word segmentation
2020-01-31 17:45:00 +01:00
