Deep Reader: Information extraction from Document images via relation extraction and Natural Language

TL;DR

DeepReader is an end-to-end framework that combines advanced vision algorithms and relational modeling to extract structured information from document images, enabling natural language querying for non-technical users.

Contribution

It introduces a novel integrated system that captures visual entities and their relationships in documents, facilitating easy information retrieval through natural language queries.

Findings

Effective recognition of handwritten and printed text.

Successful extraction of visual entities like tables and boxes.

Natural language interface enables non-technical user queries.

Abstract

Recent advancements in the area of Computer Vision with state-of-art Neural Networks has given a boost to Optical Character Recognition (OCR) accuracies. However, extracting characters/text alone is often insufficient for relevant information extraction as documents also have a visual structure that is not captured by OCR. Extracting information from tables, charts, footnotes, boxes, headings and retrieving the corresponding structured representation for the document remains a challenge and finds application in a large number of real-world use cases. In this paper, we propose a novel enterprise based end-to-end framework called DeepReader which facilitates information extraction from document images via identification of visual entities and populating a meta relational model across different entities in the document image. The model schema allows for an easy to understand abstraction of the entities detected by the deep vision models and the relationships between them. DeepReader has a suite of state-of-the-art vision algorithms which are applied to recognize handwritten and printed text, eliminate noisy effects, identify the type of documents and detect visual entities like tables, lines and boxes. Deep Reader maps the extracted entities into a rich relational schema so as to capture all the relevant relationships between entities (words, textboxes, lines etc) detected in the document. Relevant information and fields can then be extracted from the document by writing SQL queries on top of the relationship tables. A natural language based interface is added on top of the relationship schema so that a non-technical user, specifying the queries in natural language, can fetch the information with minimal effort. In this paper, we also demonstrate many different capabilities of Deep Reader and report results on a real-world use case.