1. preface

With the continuous iteration and application innovation of cutting-edge technologies such as big data, cloud computing, blockchain and artificial intelligence, financial technology has devoted to the transformation of traditional financial industry. It has become a major trend to use financial technology to improve the efficiency of existing business models and customer experience in the global business society. The automatic analysis of complex typesetting tables is an important technical branch of artificial intelligence. It can extract the bordered tables and complex typesetting tables from massive data files, which could greatly improve the work efficiency of financial analysts. The technology could apply to various financial scenarios such as intelligent investment research and risk control.

2. Product characteristics and advantages

This artificial intelligence product is developed independently by P.A.T. Tech. The main function is to extract all tables, including the bordered tables and complex typesetting tables extraction from financial documents, which can be copied and downloaded. As one of the global leaders in the table extraction field, PDFlux’s accuracy rate is over 99%. It is a world-class technical problem to extract table with complex typesetting, and there are various products or technical solutions in the market to solve it. Most of the underlying technology of these products is depended on the manual copying and pasting and semi-automatic tools, which is then verified by manual. This is the reason why existing financial investment research or data providers can only provide the limited common data and lag in the data updating. Such tables usually do not have a complete wireframe, which can only be judged by the lines, patches, and semantics. PDFlux makes full use of artificial intelligence and deep learning technology to identify the structure of complex typesetting tables, and extracts various types of tables from financial documents. It greatly reduces the loss in the unstructured data conversion and improves the validity of the tabular structured data extraction. Currently, we can provide privatization deployment and SaaS services (toB), as well as PDFlux client (toC, for customer experience).