Back to the newsroom

Press Release

Korea Deep Learning develops a deep learning OCR system with NFC&T to read mail addresses

Korea Deep LearningPublished

Korea Deep Learning develops a deep learning OCR system with NFC&T to read mail addresses
In September 2024, Korea Deep Learning is developing an OCR system with NFC&T that automatically reads the destination addresses on registered and nonstandard mail and sorts it.

In September 2024, Korea Deep Learning is developing an OCR system for reading mail addresses with its client company NFC&T, through the Data Voucher Support Program managed by the Korea Data Agency. Deep learning OCR automatically reads the destination addresses on registered and nonstandard mail and creates sorting information, replacing the sorting that workers did by hand with a logistics automation system.

From hand sorting to automatic sorting for registered and nonstandard mail

The project was planned to automate the sorting of registered and nonstandard mail, which workers currently sort by hand.

Deep learning OCR that reads addresses and creates sorting information

At the core of the system, deep learning OCR automatically reads the destination address written on the surface of each piece of mail. Based on the address it reads, the system creates sorting information and passes it on to the next logistics step. When AI takes over the step in which people read and sort addresses, it can speed up sorting and reduce missorting caused by differences in judgment between workers. Through this, Korea Deep Learning and NFC&T aim to change every stage of mail processing.

Logistics experience from Korea Post's high-speed OCR and handwritten waybill recognition

Korea Deep Learning brings to the project the experience in recognizing logistics documents that it built through Korea Post's high-speed logistics OCR and its project to digitize handwritten waybills. DEEP OCR, the character recognition technology it unveiled in 2021, was applied to Korea Post's high-speed logistics OCR system, which recognizes mail labels within 0.1 seconds. In 2022, it was selected with Sunghwa Corporate Express for the AI Voucher Support Program and developed a "high-speed deep learning waybill recognition system for digitizing handwritten waybills," and the automatic parcel waybill recognition system built at the time was expected to cut the client company's labor costs by KRW 450 million a year.

Ideas proven in the logistics field also earned outside recognition. Its "deep learning-based dual recognition system for shipping information," which instantly re-reads the address at high speed when a barcode is misread, won the grand prize in the finals of the 2022 New Gyeonggi Startup Competition in September 2022, in which 392 teams competed. Its experience processing handwritten waybills and logistics images moving at high speed is the foundation for reading mail addresses.

From Data Voucher supplier to logistics automation

The Data Voucher Support Program uses vouchers to cover the costs of companies that need to buy and process data, and Korea Deep Learning takes part in it as a supplier. In 2022, 22 projects in which it took part as a supplier were selected all at once for the Ministry of Science and ICT's Data Voucher Support Program.

A company official said that through this project, Korea Deep Learning would greatly improve efficiency in the mail and logistics industries and move ahead in the field of logistics automation systems.

About DEEP OCR

Korea Deep Learning started out by reading documents. Built on KDL's own document-specialized VLM, DEEP OCR is Korea's first VLM-based OCR solution, reading not just the characters but the structure and meaning of a document together. It handles unstructured documents in every kind of layout without building separate training data, and extracts text and key-value data from visual elements such as seals, signatures, tables, numbers and symbols. From loan documents in finance to administrative documents in the public sector and handwritten waybills in logistics, documents people used to copy out by hand become business data through DEEP OCR. More information is available on the DEEP OCR page at koreadeep.com.

About Korea Deep Learning

Korea Deep Learning (KDL), founded in 2019, is an enterprise AI agent company. Built on KDL Frontier, its own vision-language model (VLM) specialized for documents, KDL makes AI agents that take on and complete the document work of businesses and public institutions. Its main products are DEEP OCR, DEEP Parser, DEEP Agent and DEEP Ops.

Source