Current location - Education and Training Encyclopedia - Resume - OCR recognition-financial statement recognition
OCR recognition-financial statement recognition
First of all, financial report is also called financial report, including balance sheet, income statement and cash flow statement. Financial statements are of great reference significance to the operation of enterprises: they can comprehensively and systematically reveal the financial situation, operating results and cash flow of enterprises in a certain period; It is beneficial for investors, creditors and other interested parties to grasp the financial situation, operating results and cash flow of the enterprise, and then analyze the profitability, solvency, investment income and development prospects of the enterprise, and provide decision-making basis for its investment, loan and trade; Meet the supervision of finance, taxation, industry and commerce, auditing and other departments on enterprise management.

At present, the data analysis of financial statements is basically based on manual reading and auditing, which is inefficient and cannot realize the automation of examination and approval. In the part of bank credit review, the traditional manual review method has seriously hindered the efficiency and business development of bank credit review. In fact, the basis of credit review is the core data in financial statements. As long as the collection of financial statement data is improved, the efficiency of bank credit audit will be greatly improved. Next, I will introduce the OCR financial statement identification technology of Zhongan in the future, which will help banks, taxation and auditing improve their efficiency and build an automatic credit audit system.

Zhongan future? OCR? Financial report identification technology is to transform words, tables and images on images into electronic data with the help of image processing technology, and quickly collect data information through computer programs. There are a lot of data in financial statements that need to be entered manually. Using OCR to identify financial statements will greatly improve the efficiency and accuracy of data collection.

OCR? Advantages of financial report identification technology:

Fast identification of financial statements: It takes only a few seconds to identify a page of statements by using OCR identification technology;

Financial report identification collects comprehensive information: OCR identification will get all the detailed data of the enterprise;

Financial report identification technology does not need templates: OCR financial report identification software does not need users to set identification templates. Regardless of the report length, Zhongan's future financial statement identification technology can automatically analyze the matching rules and achieve the purpose of accurate identification.

Original format restoration of financial statement recognition technology: traditional OCR text recognition software can only deal with the recognition of full text and simple table structure, and some OCR software still need to make recognition templates to recognize it, which can not meet the recognition of such a complex layout of financial statements. In this financial report identification technology, we can restore a table with the same logical structure as the financial report.

Support the identification of JPG and pdf report files, and output XML and JSON identification result files.

Support automatic classification of three financial statements.

Support account dictionary matching of financial statements.

Support filtering the interference of red chapter.

Support automatic analysis of reports with table lines and repair of border lines.

Support the analysis and identification of internal reports in tables.

Support for identifying reports with complex titles.

Support the identification of left and right tables and upper and lower tables

Built-in financial cross-checking formula. If the formula is not satisfied, it will be displayed in red (an existing known formula is embedded).

When two images correspond to the same table, the recognition result is output in XML format.

Support for Windows or Linux platform deployment.

Provide http standard interface