add pagexml2imag
b-vr103
5 years ago
commit
003bfc2cd0
@ -0,0 +1,24 @@
|
||||
<?xml version="1.0" encoding="UTF-8"?>
|
||||
<projectDescription>
|
||||
<name>toolbox_pagexml2img</name>
|
||||
<comment></comment>
|
||||
<projects>
|
||||
</projects>
|
||||
<buildSpec>
|
||||
<buildCommand>
|
||||
<name>org.python.pydev.PyDevBuilder</name>
|
||||
<arguments>
|
||||
</arguments>
|
||||
</buildCommand>
|
||||
</buildSpec>
|
||||
<natures>
|
||||
<nature>org.python.pydev.pythonNature</nature>
|
||||
</natures>
|
||||
<linkedResources>
|
||||
<link>
|
||||
<name>setup.py</name>
|
||||
<type>1</type>
|
||||
<location>/home/vahid/workspace/toolbox_pagexml2img/setup.py</location>
|
||||
</link>
|
||||
</linkedResources>
|
||||
</projectDescription>
|
||||
@ -0,0 +1,8 @@
|
||||
<?xml version="1.0" encoding="UTF-8" standalone="no"?>
|
||||
<?eclipse-pydev version="1.0"?><pydev_project>
|
||||
<pydev_pathproperty name="org.python.pydev.PROJECT_SOURCE_PATH">
|
||||
<path>/${PROJECT_DIR_NAME}</path>
|
||||
</pydev_pathproperty>
|
||||
<pydev_property name="org.python.pydev.PYTHON_PROJECT_VERSION">python interpreter</pydev_property>
|
||||
<pydev_property name="org.python.pydev.PYTHON_PROJECT_INTERPRETER">Default</pydev_property>
|
||||
</pydev_project>
|
||||
@ -0,0 +1 @@
|
||||
from pagexml2img import *
|
||||
@ -0,0 +1,160 @@
|
||||
#! /usr/bin/env python3
|
||||
|
||||
__version__= '1.0'
|
||||
|
||||
import argparse
|
||||
import sys
|
||||
import os
|
||||
import numpy as np
|
||||
import warnings
|
||||
import xml.etree.ElementTree as ET
|
||||
from tqdm import tqdm
|
||||
import cv2
|
||||
|
||||
with warnings.catch_warnings():
|
||||
warnings.simplefilter("ignore")
|
||||
|
||||
__doc__=\
|
||||
"""
|
||||
tool to extract 2d or 3d RGB images from page xml data. In former case output will be 1
|
||||
2D image array which each class has filled with a pixel value. In the case of 3D RGB image
|
||||
each class will be defined with a RGB value and beside images a text file of classes also will be produced.
|
||||
This classes.txt file is required for dhsegment tool.
|
||||
"""
|
||||
|
||||
class pagexml2img:
|
||||
def __init__(self,dir_in, out_dir,output_type):
|
||||
self.dir=dir_in
|
||||
self.output_dir=out_dir
|
||||
self.output_type=output_type
|
||||
|
||||
def get_content_of_dir(self):
|
||||
"""
|
||||
Listing all ground truth page xml files. All files are needed to have xml format.
|
||||
"""
|
||||
|
||||
gt_all=os.listdir(self.dir)
|
||||
self.gt_list=[file for file in gt_all if file.split('.')[ len(file.split('.'))-1 ]=='xml' ]
|
||||
|
||||
def get_images_of_ground_truth(self):
|
||||
"""
|
||||
Reading the page xml files and write the ground truth images into given output directory.
|
||||
"""
|
||||
|
||||
if self.output_type=='3d' or self.output_type=='3D':
|
||||
classes=np.array([ [0,0,0],[255,0,0],[0,255,0],[0,0,255]])
|
||||
|
||||
|
||||
|
||||
|
||||
for index in tqdm(range(len(self.gt_list))):
|
||||
try:
|
||||
tree1 = ET.parse(self.dir+'/'+self.gt_list[index])
|
||||
root1=tree1.getroot()
|
||||
|
||||
|
||||
for jj in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Page'):
|
||||
y_len=int(jj.attrib['imageHeight'])
|
||||
x_len=int(jj.attrib['imageWidth'])
|
||||
|
||||
co_text=[]
|
||||
co_sep=[]
|
||||
co_img=[]
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}TextRegion'):
|
||||
c_t_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_t_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_text.append(np.array(c_t_in))
|
||||
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}ImageRegion'):
|
||||
c_i_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_i_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_img.append(np.array(c_i_in))
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}SeparatorRegion'):
|
||||
c_s_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_s_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_sep.append(np.array(c_s_in))
|
||||
|
||||
img = np.zeros( (y_len,x_len,3) )
|
||||
img_poly=cv2.fillPoly(img, pts =co_text, color=(255,0,0))
|
||||
img_poly=cv2.fillPoly(img, pts =co_img, color=(0,255,0))
|
||||
img_poly=cv2.fillPoly(img, pts =co_sep, color=(0,0,255))
|
||||
|
||||
try:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('-')[1].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
pass
|
||||
np.savetxt(self.output_dir+'/../classes.txt',classes)
|
||||
|
||||
if self.output_type=='2d' or self.output_type=='2D':
|
||||
|
||||
for index in tqdm(range(len(self.gt_list))):
|
||||
try:
|
||||
|
||||
tree1 = ET.parse(self.dir+'/'+self.gt_list[index])
|
||||
root1=tree1.getroot()
|
||||
|
||||
for jj in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Page'):
|
||||
y_len=int(jj.attrib['imageHeight'])
|
||||
x_len=int(jj.attrib['imageWidth'])
|
||||
|
||||
co_text=[]
|
||||
co_sep=[]
|
||||
co_img=[]
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}TextRegion'):
|
||||
c_t_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_t_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_text.append(np.array(c_t_in))
|
||||
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}ImageRegion'):
|
||||
c_i_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_i_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_img.append(np.array(c_i_in))
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}SeparatorRegion'):
|
||||
c_s_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_s_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_sep.append(np.array(c_s_in))
|
||||
|
||||
img = np.zeros( (y_len,x_len) )
|
||||
img_poly=cv2.fillPoly(img, pts =co_text, color=(1,1,1))
|
||||
img_poly=cv2.fillPoly(img, pts =co_img, color=(2,2,2))
|
||||
img_poly=cv2.fillPoly(img, pts =co_sep, color=(3,3,3))
|
||||
try:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('-')[1].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
pass
|
||||
def run(self):
|
||||
self.get_content_of_dir()
|
||||
self.get_images_of_ground_truth()
|
||||
def main():
|
||||
parser=argparse.ArgumentParser()
|
||||
|
||||
parser.add_argument('-dir_in','--dir_in', dest='inp1', default=None, help='directory of page-xml files')
|
||||
parser.add_argument('-dir_out','--dir_out', dest='inp2', default=None, help='directory where ground truth images would be written')
|
||||
parser.add_argument('-type','--type', dest='inp3', default=None, help='this defines how output should be. A 2d image array or a 3d image array encoded with RGB color. Just pass 2d or 3d. The file will be saved one directory up. 2D image array is 3d but only information of one channel would be enough since all channels have the same values.')
|
||||
options=parser.parse_args()
|
||||
|
||||
possibles=globals()
|
||||
possibles.update(locals())
|
||||
x=pagexml2img(options.inp1,options.inp2,options.inp3)
|
||||
x.run()
|
||||
if __name__=="__main__":
|
||||
main()
|
||||
|
||||
|
||||
|
||||
Binary file not shown.
@ -0,0 +1,9 @@
|
||||
all: build install
|
||||
|
||||
build:
|
||||
python3 setup.py build
|
||||
|
||||
install:
|
||||
python3 setup.py install --user
|
||||
cp pagexml2img/pagexml2img.py /home/vahid/bin/pagexml2img
|
||||
chmod +x /home/vahid/bin/pagexml2img
|
||||
@ -0,0 +1,11 @@
|
||||
Metadata-Version: 1.0
|
||||
Name: pagexml2img
|
||||
Version: 1.0
|
||||
Summary: UNKNOWN
|
||||
Home-page: UNKNOWN
|
||||
Author: UNKNOWN
|
||||
Author-email: UNKNOWN
|
||||
License: UNKNOWN
|
||||
Description-Content-Type: UNKNOWN
|
||||
Description: UNKNOWN
|
||||
Platform: UNKNOWN
|
||||
@ -0,0 +1,7 @@
|
||||
setup.py
|
||||
pagexml2img/__init__.py
|
||||
pagexml2img/pagexml2img.py
|
||||
pagexml2img.egg-info/PKG-INFO
|
||||
pagexml2img.egg-info/SOURCES.txt
|
||||
pagexml2img.egg-info/dependency_links.txt
|
||||
pagexml2img.egg-info/top_level.txt
|
||||
@ -0,0 +1 @@
|
||||
|
||||
@ -0,0 +1 @@
|
||||
pagexml2img
|
||||
@ -0,0 +1 @@
|
||||
from pagexml2img import *
|
||||
@ -0,0 +1,160 @@
|
||||
#! /usr/bin/env python3
|
||||
|
||||
__version__= '1.0'
|
||||
|
||||
import argparse
|
||||
import sys
|
||||
import os
|
||||
import numpy as np
|
||||
import warnings
|
||||
import xml.etree.ElementTree as ET
|
||||
from tqdm import tqdm
|
||||
import cv2
|
||||
|
||||
with warnings.catch_warnings():
|
||||
warnings.simplefilter("ignore")
|
||||
|
||||
__doc__=\
|
||||
"""
|
||||
tool to extract 2d or 3d RGB images from page xml data. In former case output will be 1
|
||||
2D image array which each class has filled with a pixel value. In the case of 3D RGB image
|
||||
each class will be defined with a RGB value and beside images a text file of classes also will be produced.
|
||||
This classes.txt file is required for dhsegment tool.
|
||||
"""
|
||||
|
||||
class pagexml2img:
|
||||
def __init__(self,dir_in, out_dir,output_type):
|
||||
self.dir=dir_in
|
||||
self.output_dir=out_dir
|
||||
self.output_type=output_type
|
||||
|
||||
def get_content_of_dir(self):
|
||||
"""
|
||||
Listing all ground truth page xml files. All files are needed to have xml format.
|
||||
"""
|
||||
|
||||
gt_all=os.listdir(self.dir)
|
||||
self.gt_list=[file for file in gt_all if file.split('.')[ len(file.split('.'))-1 ]=='xml' ]
|
||||
|
||||
def get_images_of_ground_truth(self):
|
||||
"""
|
||||
Reading the page xml files and write the ground truth images into given output directory.
|
||||
"""
|
||||
|
||||
if self.output_type=='3d' or self.output_type=='3D':
|
||||
classes=np.array([ [0,0,0],[255,0,0],[0,255,0],[0,0,255]])
|
||||
|
||||
|
||||
|
||||
|
||||
for index in tqdm(range(len(self.gt_list))):
|
||||
try:
|
||||
tree1 = ET.parse(self.dir+'/'+self.gt_list[index])
|
||||
root1=tree1.getroot()
|
||||
|
||||
|
||||
for jj in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Page'):
|
||||
y_len=int(jj.attrib['imageHeight'])
|
||||
x_len=int(jj.attrib['imageWidth'])
|
||||
|
||||
co_text=[]
|
||||
co_sep=[]
|
||||
co_img=[]
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}TextRegion'):
|
||||
c_t_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_t_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_text.append(np.array(c_t_in))
|
||||
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}ImageRegion'):
|
||||
c_i_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_i_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_img.append(np.array(c_i_in))
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}SeparatorRegion'):
|
||||
c_s_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_s_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_sep.append(np.array(c_s_in))
|
||||
|
||||
img = np.zeros( (y_len,x_len,3) )
|
||||
img_poly=cv2.fillPoly(img, pts =co_text, color=(255,0,0))
|
||||
img_poly=cv2.fillPoly(img, pts =co_img, color=(0,255,0))
|
||||
img_poly=cv2.fillPoly(img, pts =co_sep, color=(0,0,255))
|
||||
|
||||
try:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('-')[1].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
pass
|
||||
np.savetxt(self.output_dir+'/../classes.txt',classes)
|
||||
|
||||
if self.output_type=='2d' or self.output_type=='2D':
|
||||
|
||||
for index in tqdm(range(len(self.gt_list))):
|
||||
try:
|
||||
|
||||
tree1 = ET.parse(self.dir+'/'+self.gt_list[index])
|
||||
root1=tree1.getroot()
|
||||
|
||||
for jj in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Page'):
|
||||
y_len=int(jj.attrib['imageHeight'])
|
||||
x_len=int(jj.attrib['imageWidth'])
|
||||
|
||||
co_text=[]
|
||||
co_sep=[]
|
||||
co_img=[]
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}TextRegion'):
|
||||
c_t_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_t_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_text.append(np.array(c_t_in))
|
||||
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}ImageRegion'):
|
||||
c_i_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_i_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_img.append(np.array(c_i_in))
|
||||
|
||||
for nn in root1.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}SeparatorRegion'):
|
||||
c_s_in=[]
|
||||
for ll in nn.iter('{http://schema.primaresearch.org/PAGE/gts/pagecontent/2010-03-19}Point'):
|
||||
c_s_in.append([ int(np.float(ll.attrib['x'])) , int(np.float(ll.attrib['y'])) ])
|
||||
co_sep.append(np.array(c_s_in))
|
||||
|
||||
img = np.zeros( (y_len,x_len) )
|
||||
img_poly=cv2.fillPoly(img, pts =co_text, color=(1,1,1))
|
||||
img_poly=cv2.fillPoly(img, pts =co_img, color=(2,2,2))
|
||||
img_poly=cv2.fillPoly(img, pts =co_sep, color=(3,3,3))
|
||||
try:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('-')[1].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
cv2.imwrite(self.output_dir+'/'+self.gt_list[index].split('.')[0]+'.png',img_poly )
|
||||
except:
|
||||
pass
|
||||
def run(self):
|
||||
self.get_content_of_dir()
|
||||
self.get_images_of_ground_truth()
|
||||
def main():
|
||||
parser=argparse.ArgumentParser()
|
||||
|
||||
parser.add_argument('-dir_in','--dir_in', dest='inp1', default=None, help='directory of page-xml files')
|
||||
parser.add_argument('-dir_out','--dir_out', dest='inp2', default=None, help='directory where ground truth images would be written')
|
||||
parser.add_argument('-type','--type', dest='inp3', default=None, help='this defines how output should be. A 2d image array or a 3d image array encoded with RGB color. Just pass 2d or 3d. The file will be saved one directory up. 2D image array is 3d but only information of one channel would be enough since all channels have the same values.')
|
||||
options=parser.parse_args()
|
||||
|
||||
possibles=globals()
|
||||
possibles.update(locals())
|
||||
x=pagexml2img(options.inp1,options.inp2,options.inp3)
|
||||
x.run()
|
||||
if __name__=="__main__":
|
||||
main()
|
||||
|
||||
|
||||
|
||||
Loading…
Reference in New Issue