Home case Finance Wintone OCR improves financial credit approval efficiency

Wintone OCR improves financial credit approval efficiency

2022-10-09 09:59:15

As various regions ramp up financial support for economic recovery, they also place significant pressure on banks and financial institutions' credit approval processes. In actual credit business and financial institution loan approvals, customer qualifications must be verified, requiring submission of multiple documents such as financial reports, property certificates, business licenses, and bank cards. The manual review process is time-consuming, error-prone, and significantly slows down the credit approval process.

Wintone OCR Recognition Service

Wintone Technology's OCR recognition service enables full-field recognition and intelligent data entry of various credit review documents, eliminating repetitive manual data entry tasks. After recognition, the system outputs structured data, which is returned to the business system, standardizing the approval process, improving operational efficiency, and enhancing customer experience.

01 Financial Report Recognition

Wintone's financial report recognition system is an OCR-based solution developed specifically for financial industry risk control data entry. It supports automatic recognition and classification of the balance sheet, cash flow statement, and income statement.

wintone ocr

It supports importing files in formats such as JPG, BMP, TIF, PDF, EXCEL, etc. The output formats of recognition results include XML, JSON, or EXCEL, and standard HTTP or WebService interfaces are provided for seamless integration with credit systems and other business systems.

It supports subtotal validation, total validation, and accounting equation checks, and uses data validation and confidence analysis to highlight disputed data. It also supports automatic matching of various report items.

The recognition accuracy for individual fields is over 95%, and the recognition speed for a single report is around 3 seconds.

The system provides templates for new and old accounting standards released by the Ministry of Finance, and supports online configuration for users to map specific report modules with fuzzy mapping.

The system can quickly and accurately extract financial statement data from hundreds of audit reports, supporting the recognition of reports with no borders, incomplete borders, or bordered formats. It also supports red stamp detection, noise removal, and multi-angle rotation for image processing.

02 Real Estate Certificate/Property Title Recognition

Wintone's real estate certificate recognition system can quickly and accurately recognize various real estate certificate information, supporting full-field recognition and output of property rights or matters, property owners, obligors, and other information with a response time of seconds.

Wintone's property title recognition system supports the recognition of certificate number, property location, property nature, registration time, owner, and other key fields.

The system supports recognizing various versions of real estate certificates, including the 2008 edition, 2000 edition, Guangdong 2008 edition, Shanghai 2000 edition, Shenzhen 2005 edition, Tianjin edition, Qingdao property title, and others. It intelligently classifies them to meet the credit review requirements of banks and financial institutions nationwide.

03 Bank Card Recognition

Wintone's bank card recognition system supports the recognition of full-field information such as card number, type, expiration date, name (in pinyin), and validity period. It supports recognition of flat cards, embossed cards, vertical cards, irregular cards, and international cards. The system includes advanced image processing functions to handle difficult samples, such as those with large backgrounds, uneven lighting, or large angles. It supports automatic rotation and recognition of images at 45°, 90°, 180°, and 270° angles.

04 Business License Recognition

Wintone's business license recognition system supports the recognition of various styles, including the "Three-in-One" and "Five-in-One" certificates, as well as both horizontal and vertical formats. It extracts and outputs key information such as the unified social credit code, company name, and address, and supports recognition of rare characters. Regardless of blurry images, large-angle images, or complex backgrounds, the system can intelligently detect and accurately recognize the data.

Additionally, Wintone Technology's document recognition services, including ID card recognition, invoice recognition, and contract recognition, contribute to the credit approval process. Wintone’s OCR recognition service has been widely applied in credit approval scenarios for financial institutions, banks, and other sectors, including companies like Lakala, JD Finance, and Tianjin Housing Provident Fund Center, helping to reduce costs, improve efficiency, and enhance service quality.