SYSTEM AND METHOD FOR EXTRACTING TABULAR DATA FROM A DOCUMENT

Patent №

US 11,048,867

Granted

2021-06-29

Filed 2019

Owner

WIPRO LIMITED

Lab

AI components

3

nlp · vision · planning

Assignment

Recorded

Dataset

AIPD

2023_r1 edition

Application

16674054

The present invention relates to a method for extracting tabular data from a document. The method includes identifying a bordered table or a borderless table in a received document and an image of the document. The tabular data in the identified bordered table is extracted using a first and a second set of pixel coordinates from the plurality of pixel coordinates. Further, upon identifying the borderless table in the document, a first set of document coordinates of at least one row of the borderless table is determined. Furthermore, a second set of document coordinates of the at least one column corresponding to the at least one row is determined. Finally, the tabular data in the identified borderless table is extracted from the document based on the determined first and second set of document coordinates.

Natural languageVisionPlanningG06F 40/177G06F 40/103G06V 10/82G06V 30/19173G06V 30/412G06V 30/413G06V 30/414

AI classification

Vision1.00
Natural language0.96
Planning0.57
Knowledge representation0.21
Machine learning0.04
AI hardware0.03
Evolutionary computation0.00
Speech0.00

Ownership

WIPRO LIMITED

assignment · 509120711

Assignors

AGGARWAL, AMIT, YELLAPRAGADA, KRISHNA PRASAD, URS, VEENA SRIKANTH RAJE

On an employer assignment, the assignors are typically the inventors.

From the same owner

© 2026 NYSGPT2525 LLC