<s>
Table	B-General_Concept
extraction	I-General_Concept
is	O
the	O
process	O
of	O
recognizing	O
and	O
separating	O
a	O
table	B-Application
from	O
a	O
large	O
document	O
,	O
possibly	O
also	O
recognizing	O
individual	O
rows	O
,	O
columns	O
or	O
elements	O
.	O
</s>
<s>
It	O
may	O
be	O
regarded	O
as	O
a	O
special	O
form	O
of	O
information	B-General_Concept
extraction	I-General_Concept
.	O
</s>
<s>
Table	B-General_Concept
extractions	I-General_Concept
from	O
webpages	O
can	O
take	O
advantage	O
of	O
the	O
special	O
HTML	B-General_Concept
elements	I-General_Concept
that	O
exist	O
for	O
tables	O
,	O
e.g.	O
,	O
the	O
"	O
table	B-Application
"	O
tag	B-General_Concept
,	O
</s>
<s>
and	O
programming	O
libraries	O
may	O
implement	O
table	B-General_Concept
extraction	I-General_Concept
from	O
webpages	O
.	O
</s>
<s>
The	O
Python	B-Application
pandas	I-Application
software	O
library	O
can	O
extract	O
tables	O
from	O
HTML	O
webpages	O
via	O
its	O
read_html( )	O
function	O
.	O
</s>
<s>
More	O
challenging	O
is	O
table	B-General_Concept
extraction	I-General_Concept
from	O
PDFs	B-Application
or	O
scanned	B-Application
images	I-Application
,	O
where	O
there	O
usually	O
is	O
no	O
table-specific	O
machine	O
readable	O
markup	O
.	O
</s>
<s>
Systems	O
that	O
extract	O
data	O
from	O
tables	O
in	O
scientific	O
PDFs	B-Application
have	O
been	O
described	O
.	O
</s>
<s>
Large-scale	O
table	B-General_Concept
extraction	I-General_Concept
of	O
Wikipedia	O
infoboxes	O
forms	O
one	O
of	O
the	O
sources	O
for	O
DBpedia	B-Application
.	O
</s>
<s>
Commercial	O
web	O
services	O
for	O
table	B-General_Concept
extraction	I-General_Concept
exist	O
,	O
e.g.	O
,	O
Amazon	B-Application
Textract	O
,	O
Google	B-Application
's	I-Application
Document	B-Application
AI	I-Application
,	O
IBM	O
Watson	O
Discovery	O
,	O
and	O
Microsoft	O
Form	O
Recognizer	O
.	O
</s>
<s>
Open	O
source	O
tools	O
also	O
exist	O
,	O
e.g.	O
,	O
PDFFigures	O
2.0	O
that	O
has	O
been	O
used	O
in	O
Semantic	B-Application
Scholar	I-Application
.	O
</s>
<s>
In	O
a	O
comparison	O
published	O
in	O
2017	O
,	O
the	O
researchers	O
found	O
the	O
proprietary	O
program	O
ABBYY	B-Operating_System
FineReader	I-Operating_System
to	O
yield	O
the	O
best	O
PDF	B-Application
table	B-General_Concept
extraction	I-General_Concept
performance	O
among	O
six	O
different	O
tools	O
evaluated	O
.	O
</s>
