Back to Customer Stories
Hardware & Software / Content Capture/Document Processing

ABBYY FineReader Engine Facilitates Legal Decision Making for Chinese Procurators

main image customer story YDR

Customer Overview

Name YDR (Sichuan Yadaer Technology Co., Ltd.)
Headquarters Chengdu, China
Industry Hardware & Software
Web
Challenge
  • Automate image & text document processing, while improving accuracy of data recording;
  • Reduce manual task related to document handling;
  • Reduce document processing time
Solution

Award-winning ABBYY Content Intelligence solution ABBYY FineReader Engine

Results
  • Automated dossier creation process;
  • Higher accuracy of entered records;
  • Decreased time for entering and analyzing the data

Sichuan Yadaer Technology Co. Ltd. was established in 2013. From the beginning, the enterprise has been focused on offering integrated digital solutions to government and the legal sector. One of the company’s main products is “Electronic Court Case Management System”. By identifying, accessing and linking relevant case data, this application helps to process and store legal documentation of regional Procurators and Police Bureaus. More than 500 Chinese procurators’ offices have successfully implemented Yadaer Solution.

Challenge

To ensure information is being used effectively and correctly and in order to facilitate decision-making during litigation, traditional paper files have to be replaced by electronic ones. This allows litigators to access and prepare the required data in a timely manner while working on a court case — from finding a precedent case to archiving the final documentation.

One of the features of Yadaer Solution is the creation of an electronic catalogue with all previous cases. Judges use it as precedent while working with evidence and passing the judgment. However, the creation of an electronic catalogue is a challenging task as it requires digitizing a massive amount of low quality paper case archives. The development of such a catalogue requires the usage of solution that provides outstanding OCR quality.

The current Sichuan Yadaer solution is designed to process document volumes ranging from 2000 to 5000 pages per day on average. The files use Chinese Simplified characters, as well as the Latin alphabet and numbers. Creating an electronic document catalogue and classifying a large number of files turned out to be tedious work, as all the steps in the process had to be performed manually. For instance, after categorizing files, all the information in the catalogue had to be filled in by hand along with classification information.

The amount of data was exceedingly high, so the staff often made typos or chose the wrong file directory. As a result, the catalogue information was incomplete and needed corrections to meet the file creation requirements. The variety of document types made the speed of data entry particularly low. Furthermore, given that it was paramount to ensure 100% consistency of entered data with information from the corresponding paper files, error avoidance was another major task to be addressed. This big challenge can be best tackled by introducing a solution with AI-powered OCR capabilities.

Solution

Integrating award-winning ABBYY Content Intelligence solutions into the existing solution extends the system’s capabilities and allows it to turn inaccessible information into on-demand available and actionable data. For example, the system can then compare all the information provided for a new case including witness testimonies, documents, reference materials, and extract conflicting information. Moreover, when making a decision, the judge can be presented with important facts such as the defendant’s age and previous convictions, the amount of losses caused, among other vital details.

Using Yadaer Solution algorithms empowered with ABBYY Content IQ tools to scan millions of court records and find similar cases. This in turn allows the solution to put forward a recommendation focusing on critical aspects concerning fines, imprisonment or other sentences. This scrutinizing process ensures consistency and integrity in the work of the judicial system. Recommendations derived from information analysis help to avoid bias that even experienced judges may sometimes exhibit.

To tackle the challenges associated with slow and inaccurate data entry, Yadaer chose ABBYY FineReader Engine to integrate into their full solution. Thanks to its superior OCR technology, the content of catalogue pages is now converted into files in readable XML format, right after the image is selected, scanned, and recognized. With the help of the rules that set up the conditions for extracting structured data, the content of specified fields is returned directly inside the application display window. Subsequently, the files are verified once again and are made available and ready-to-use for the staff.

The time spent entering the data has been significantly reduced. Previously, the time needed for processing all the files took 50% of the entire worktime. Overall, work efficiency has been improved as a result of two main factors: enhanced accuracy of entered records and faster rate of document processing.

Result

Thanks to ABBYY FineReader Engine, the process of text recognition is now fully automated. Yadaer Electronic Case Court Management System was developed, tested, and put into work within less than a year.

The main two goals of implementing the software have been accomplished. First the amount of manual work has been reduced, eliminating the tedious process of files classification. Second, high accuracy of information has been achieved. Yadaer has noted the exceptional quality of the developed solution, especially the high recognition rate.

“The amount of labor costs and constant need of data accuracy made us adopt new technologies. We had tried other OCR technologies previously, but then found out that they were not able to meet our recognition requirements. So, then we chose the internationally renowned ABBYY and their OCR technology. Now we can say, that after using ABBYY technology, after an initial period of optimization and operation, the recognition rate equals 95%. Such results are far beyond our expectations. We are now fully satisfied with the stable functioning and high efficiency of the entire software.”
Tang Hu, Project Manager, Yadaer