qurator-spk/sbb_pixelwise_segmentation
1
0
Fork 0
mirror of https://github.com/qurator-spk/sbb_pixelwise_segmentation.git synced 2025-06-09 20:00:05 +02:00
Code Issues Releases Wiki Activity

layout visualization updated

Browse source
This commit is contained in:
vahidrezanezhad 2025-06-02 18:23:34 +02:00
parent 13c6d28aaa
commit 834cf19cb9
2 changed files with 11 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

4
generate_gt_for_training.py
View file

@ -418,7 +418,7 @@ def visualize_textline_segmentation(dir_xml, dir_out, dir_imgs):
help="directory of images where textline segmentation will be overlayed", ) help="directory of images where textline segmentation will be overlayed", )
def visualize_layout_segmentation(xml_file, dir_xml, dir_out, dir_imgs): def visualize_layout_segmentation(xml_file, dir_xml, dir_out, dir_imgs):
assert xml_file and dir_xml, "A single xml file -xml or a dir of xml files -dx is required not both of them" assert xml_file or dir_xml, "A single xml file -xml or a dir of xml files -dx is required not both of them"
if dir_xml: if dir_xml:
xml_files_ind = os.listdir(dir_xml) xml_files_ind = os.listdir(dir_xml)
else: else:
@ -442,7 +442,7 @@ def visualize_layout_segmentation(xml_file, dir_xml, dir_out, dir_imgs):
co_text, co_graphic, co_sep, co_img, co_table, co_noise, y_len, x_len = get_layout_contours_for_visualization(xml_file) co_text, co_graphic, co_sep, co_img, co_table, co_noise, y_len, x_len = get_layout_contours_for_visualization(xml_file)
added_image = visualize_image_from_contours_layout(co_text['paragraph'], co_text['header'], co_text['drop-capital'], co_sep, co_img, co_text['marginalia'], img) added_image = visualize_image_from_contours_layout(co_text['paragraph'], co_text['header']+co_text['heading'], co_text['drop-capital'], co_sep, co_img, co_text['marginalia'], img)
cv2.imwrite(os.path.join(dir_out, f_name+'.png'), added_image) cv2.imwrite(os.path.join(dir_out, f_name+'.png'), added_image)

9
gt_gen_utils.py
View file

@ -306,6 +306,7 @@ def get_layout_contours_for_visualization(xml_file):
co_noise=[] co_noise=[]
types_text = [] types_text = []
types_graphic = []
for tag in region_tags: for tag in region_tags:
if tag.endswith('}TextRegion') or tag.endswith('}Textregion'): if tag.endswith('}TextRegion') or tag.endswith('}Textregion'):
@ -325,6 +326,9 @@ def get_layout_contours_for_visualization(xml_file):
if len(types_text_without_paragraph) == 0: if len(types_text_without_paragraph) == 0:
if "type" in nn.attrib: if "type" in nn.attrib:
c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) ) c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
else:
c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
elif len(types_text_without_paragraph) >= 1: elif len(types_text_without_paragraph) >= 1:
if "type" in nn.attrib: if "type" in nn.attrib:
if nn.attrib['type'] in types_text_without_paragraph: if nn.attrib['type'] in types_text_without_paragraph:
@ -332,10 +336,15 @@ def get_layout_contours_for_visualization(xml_file):
else: else:
c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) ) c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
else:
c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
else: else:
if "type" in nn.attrib: if "type" in nn.attrib:
if nn.attrib['type'] in all_defined_textregion_types: if nn.attrib['type'] in all_defined_textregion_types:
c_t_in[nn.attrib['type']].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) ) c_t_in[nn.attrib['type']].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
else:
c_t_in['paragraph'].append( np.array( [ [ int(x.split(',')[0]) , int(x.split(',')[1]) ] for x in p_h] ) )
break break
else: else: