Commit graph

835 commits

Author SHA1 Message Date
kba
aac2e58b05 Merge remote-tracking branch 'michalbubula/add-feedback' into prepare-release-v0.5.0 2025-09-23 19:38:56 +02:00
kba
df8d93dbfa Merge branch 'main' into add-feedback 2025-09-23 19:20:20 +02:00
kba
89e49f46bb Merge remote-tracking branch 'origin/updating_readme_for_eynollah_use_cases' into prepare-release-v0.5.0 2025-09-23 19:16:54 +02:00
vahidrezanezhad
4c2e15aa00 default cnn-rnn and transformer ocr models have changed to model_eynollah_ocr_cnnrnn_20250904 and model_eynollah_ocr_trocr_20250919 respectively 2025-09-22 13:55:31 +02:00
vahidrezanezhad
c7ed7a30e6 extending image types 2025-09-21 02:32:40 +02:00
vahidrezanezhad
eb0b44b25c Merge text of textlines and handle hyphenated words by joining them correctly 2025-09-19 23:23:30 +02:00
vahidrezanezhad
eb322d5182 writing page contour correctly in xml output + ignore unsupported file types when loading images 2025-09-19 18:06:18 +02:00
vahidrezanezhad
ee040c7767 debug new page extraction in the case of ignoring page extraction 2025-09-19 15:24:34 +02:00
kba
5c9cf8472b remove redundant/brittle interval logging 2025-09-18 13:19:57 +02:00
kba
146102842a convert all print stmts to logger.info calls 2025-09-18 13:15:18 +02:00
kba
c64d102613 move logging to CLI and make initialization optional 2025-09-18 13:07:41 +02:00
vahidrezanezhad
54d9916f3b page extraction model name is changed 2025-09-16 14:27:15 +02:00
vahidrezanezhad
52cb0d9fac new page extraction model integration 2025-09-15 13:38:23 +02:00
vahidrezanezhad
6e008345a0 new page extraction model integration 2025-09-15 13:36:58 +02:00
vahidrezanezhad
8c949cec71 PR #173 has been reverted. Additionally, for TrOCR, the cropped text lines will no longer be added to a list before prediction. Instead, for each batch size, the text line images will be collected and predictions will be made directly on them. 2025-09-03 19:18:11 +02:00
vahidrezanezhad
d9ae7bd12c merged pr #173 in #175 2025-09-02 15:27:19 +02:00
Robert Sachunsky
b84d945b5a
Merge pull request #3 from bertsky/polygon-dilate-buffer-refactor2
some refactoring (second attempt)...
2025-09-02 13:26:52 +02:00
vahidrezanezhad
92a7c7cfea changed the drop capitals bonding box to contour ratio threshold 2025-09-01 11:37:22 +02:00
vahidrezanezhad
6a735daa60
Update README.md 2025-08-31 23:30:54 +02:00
Robert Sachunsky
090341241e writer: use @type='heading' instead of 'header' 2025-08-29 17:21:30 +02:00
Robert Sachunsky
bb9cba1fd9 writer: SeparatorRegion needs SeparatorRegionType (not ImageRegionType) 2025-08-29 17:21:30 +02:00
Robert Sachunsky
eae1303ebb contours: rename 'pixel' → 'label' for clarity 2025-08-29 17:21:30 +02:00
Robert Sachunsky
dbbf1073df avoid pulling unused 'image_page_rotated' through functions 2025-08-29 17:21:30 +02:00
Robert Sachunsky
142ac8825e use box2rect instead of crop_image_inside_box when no image needed 2025-08-29 17:21:30 +02:00
Robert Sachunsky
892ff41e38 utils: introduce box2rect and box2slice 2025-08-29 17:21:30 +02:00
Robert Sachunsky
d3566e55ef polygon2contour: fix 698f38e4 (deprecated dtype) 2025-08-29 17:21:08 +02:00
Robert Sachunsky
741aa7867c get_marginals: exit early if no peaks found to avoid spurious overlap mask 2025-08-29 12:46:19 +02:00
Robert Sachunsky
57821662b9 filter_contours_without_textline_inside: avoid removing from identical lists twice 2025-08-29 12:46:12 +02:00
Robert Sachunsky
698f38e461 polygon2contour: avoid overflow 2025-08-29 12:43:46 +02:00
vahidrezanezhad
fdcae8dd6e eynollah ocr: support using either a specific model name or a models directory (default model) 2025-08-28 11:30:59 +02:00
vahidrezanezhad
7dd281267d Marginals are divided into left and right, and written from top to bottom. 2025-08-26 22:38:03 +02:00
Robert Sachunsky
fd6a6495a2 increase dilatation: textregions/lines (5→6), seplines (0→1) 2025-08-21 13:00:31 +02:00
Robert Sachunsky
8be52fb143 refactor shapely converisons into contour2polygon / polygon2contour, also handle heterogeneous geometries 2025-08-21 12:59:03 +02:00
Robert Sachunsky
8b5f90e243 move dilate_*_contours to .utils.contour, rename dilate_textregions_contours_textline_version → dilate_textline_contours 2025-08-21 01:42:46 +02:00
Robert Sachunsky
244772f086 filter_contours_area_of_image*: also ensure validity here 2025-08-21 01:33:16 +02:00
Robert Sachunsky
42474afa4b rename *lines_xml → *seplines for clarity 2025-08-21 01:32:32 +02:00
Robert Sachunsky
b610fe07a6 check_any_text_region_in_model_one_is_main_or_header_light: return original instead of resampled contours 2025-08-21 01:05:15 +02:00
Robert Sachunsky
3d53070b90 avoid creating invalid polygons via rounding 2025-08-21 01:03:46 +02:00
Robert Sachunsky
277d00579e get_textregion_contours_in_org_image_light: no back rotation, drop slope_first (always 0) 2025-08-20 14:28:14 +02:00
Robert Sachunsky
b6d1c43a85 dilate_textregions_contours_textline_version: simplify (via shapely's Polygon.buffer()), ensure validity 2025-08-20 14:26:14 +02:00
Robert Sachunsky
6c442c9ae9 separate_lines/do_work_of_slopes: skip if crop is empty 2025-08-19 22:56:36 +02:00
Robert Sachunsky
e9a6ff5d81 return_boxes_of_images_by_order_of_reading_new: simplify, avoid changing dtype during np.append 2025-08-19 20:09:09 +02:00
Robert Sachunsky
f994ea5f0b dilate_textregions_contours: simplify (via shapely's Polygon.buffer()), ensure validity 2025-08-19 11:59:26 +02:00
vahidrezanezhad
8dc2fab9fa reading order on given layout 2025-08-18 02:31:13 +02:00
Clemens Neudecker
a2359ea4c4
Merge pull request #171 from bertsky/ocrd-machine-based-ro
OCR-D processor: expose reading_order_machine_based
2025-08-15 18:40:13 +02:00
Robert Sachunsky
21615a986d OCR-D processor: expose reading_order_machine_based 2025-08-13 14:14:37 +02:00
michalbubula
8ebba5ac04 add feedback to command line interface 2025-08-12 16:21:15 +02:00
vahidrezanezhad
20614d1678 avoiding float in range 2025-08-12 12:50:15 +02:00
vahidrezanezhad
5db3e9fa64 deskewing with faster multiprocessing 2025-08-08 11:32:02 +02:00
vahidrezanezhad
a0c19c57be use the latest ocr model with balanced fraktur-antiqua training dataset 2025-08-05 14:22:22 +02:00