qurator-spk/sbb_textline_detection
1
0
Fork 0
mirror of https://github.com/qurator-spk/sbb_textline_detection.git synced 2025-10-30 17:04:16 +01:00
Code Issues Releases Wiki Activity

🐛 sbb_textline_detector: Re-base Border coords too

Browse source
This commit is contained in:
Gerber, Mike 2020-10-15 20:47:45 +02:00
parent 006c7765b0
commit 665b739fb8
1 changed files with 15 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

23
qurator/sbb_textline_detector/ocrd_cli.py
View file

@ -87,33 +87,40 @@ class OcrdSbbTextlineDetectorRecognize(Processor):
page = pcgts.get_Page() page = pcgts.get_Page()
# Merge results → PAGE file # Merge results → PAGE file
# 1. Border
if page.get_Border(): if page.get_Border():
log.warning("Page already contained a border") log.warning("Page already contained a border")
page.set_Border(tmp_page.get_Border()) # We need to translate the coordinates:
text_border = tmp_page.get_Border()
coords = text_border.get_Coords().get_points()
polygon = polygon_from_points(coords)
polygon_new = coordinates_for_segment(polygon, page_image, page_coords)
points_new = points_from_polygon(polygon_new)
coords_new = CoordsType(points=points_new)
text_border.set_Coords(coords_new)
page.set_Border(text_border)
# 2. ReadingOrder
if page.get_ReadingOrder(): if page.get_ReadingOrder():
log.warning("Page already contained a reading order") log.warning("Page already contained a reading order")
page.set_ReadingOrder(tmp_page.get_ReadingOrder()) page.set_ReadingOrder(tmp_page.get_ReadingOrder())
# 3. TextRegion
if page.get_TextRegion(): if page.get_TextRegion():
log.warning("Page already contained text regions") log.warning("Page already contained text regions")
# We need to translate the coordinates:
# We need to translate the coordinates in case we deal with a
# cropped image:
text_regions_new = [] text_regions_new = []
for text_region in tmp_page.get_TextRegion(): for text_region in tmp_page.get_TextRegion():
coords = text_region.get_Coords().get_points() coords = text_region.get_Coords().get_points()
polygon = polygon_from_points(coords) polygon = polygon_from_points(coords)
polygon_new = coordinates_for_segment(polygon, page_image, page_coords) polygon_new = coordinates_for_segment(polygon, page_image, page_coords)
points_new = points_from_polygon(polygon_new) points_new = points_from_polygon(polygon_new)
coords_new = CoordsType(points=points_new) coords_new = CoordsType(points=points_new)
text_region.set_Coords(coords_new) text_region.set_Coords(coords_new)
text_regions_new.append(text_region) text_regions_new.append(text_region)
page.set_TextRegion(text_regions_new) page.set_TextRegion(text_regions_new)
# Save metadata about this operation # Save metadata about this operation
metadata = pcgts.get_Metadata() metadata = pcgts.get_Metadata()
metadata.add_MetadataItem( metadata.add_MetadataItem(