The company is an independent body established to render professional services to prominent fund house, automated trading service providers, brokers and intermediaries etc., to meet corresponding compliance and regulations, mitigate systematic risks, reviews and alerts of companies’ activities that may cause misconduct. Through these services, companies can maintain financial soundness and operate in high standard to be competitive in the investment domain of Hong Kong Special Administrative Region.
The operational activities require fast and accurate processing of a large number of documents from different sources. This includes the digitizing of physical documents, reviewing document submissions and text searching. The project aims to enable one centralized system serving multiple internal departments, with an easy management console for system alert, traffic review and usage reporting.
The documents vary from papers and articles to forms, charts and even entire books often in mixed languages: Chinese (Traditional / Simplified) and English. When searching for a new solution, the company pursued a number of objectives. One of them was to reduce the file size after applying OCR. Since some documents contain more than 1000 pages, the download time and the bandwidth had to be reduced to minimum. Another objective was to centralize document processing and to provide transparent reporting. Finally, the aim was to increase the processing speed and the capacity as the volume of incoming documents significantly increases each year.
The challenge was well in the strengths of Nikoyo (HK) Limited; Nikoyo is a prominent system integrator and an authorized ABBYY partner in Hong Kong.
The solution choice was based on four main criteria: OCR accuracy, high-quality file-size compression, scalability and flexibility of integration and workflow configuration. Nikoyo developed a customized Web Management console driving the Recognition Server in the backend. The rich API of ABBYY Recognition Server made this development an easy task.
With ABBYY Recognition Server, the workflow process starts with the digitization of incoming documents like forms, charts, articles and books. The system module responsible for import receives images and documents from multi-functional copiers, faxes, emails, web portals and other sources and automatically directs them to Recognition Server for the document conversion. Before applying the optical character recognition (OCR), the quality of received images is enhanced for better results.
The solution provides high speed and accurate recognition of Asian languages as well as the mixed combinations with European languages. The solution runs on a single Virtual Machine Server with 12 cores. Overall, the process time is shortened significantly when compared to the legacy OCR system. The fully searchable documents are saved as PDF compressed files. The PDF format is used in electronic archives for data storage, where small file size with high visual data quality are of significant importance, just like in this case. A special compression technology called MRC (Mixed Raster Content) is used to minimize the size of PDF and PDF/A files. The MRC technology allows for better document appearance and up to 8–10 smaller file size than JPEG.
At the final stage, the captured and transformed data is sent to the corporate DMS, network folder or to a specific e-mail. The solution helps the company optimize business processes while achieving compliance, collaboration and long-term storage goals.
The project was delivered in 5 months including 3 months for development and 2 months for user testing.
The new corporate-level platform providing transparent document processing has already shown great performance. The output PDF files are reasonably compressed and fully accessible — without the time-consuming download of heavy archives. The IT support work is much smoother than before thanks to simple deployment and easy maintenance. The advanced monitoring and reporting console helps to split transparently the consumption and expenses.
ABBYY Recognition Server helped to create a reliable centralized document processing platform with an easy and transparent management console for system alert, traffic review and usage reporting. The solution has significantly reduced time and resources for support required from our IT department. The company plans to connect all the internal departments to the centralized document-conversion platform ramping up the volume to at least 3 million documents on a yearly base. Besides, the company will also organize the export of legal files from the DMS system to the solution for compressing and recording, as there is an archive of 3 million documents pending re-processing.