Brief History

An OCR device scans paper media optically, extracts an image, and identifies and recognizes characters in the image. From the standpoint of the connected computer, an OCR device is a character input device that takes the place of a keyboard. From the standpoint of the user, an OCR device is a labor-saving equipment that can automatically input large amounts of text without the manual labor of a typist using a keyboard.

Internationally, the development of character recognition technology started in before or after of 1900, and the first OCR product that read printed characters got available around 1950. The first products of domestic OCR devices were motivated by the postal code concept that the Ministry of Posts and Telecommunications at that time introduced as part of the postal service automation. The first commercially available products were Toshiba’s TR-3 and TR-4 mail sorting machines in July 1968. These products recognized free handwritten numbers for the first time in the world. By reading postal codes written in red boxes printed upper side on postcards, these products realized the“dispatch division”scheme, which sorted mail items by the positions of post offices that provided the collection and delivery service. This scheme was followed by NEC to provide the automatic postal code sorting machine of NAS-5B in June 1969. In the 1980s, the NAS-80 mail sorting machine made it possible to read printed postal codes so that mail items with either printed or handwritten postal codes could be sorted together. In 1989, the TR-17 mail sorting machine came out with the capability to recognize addresses written with kanji characters, for the first time in the world. This model switched from the dispatch division scheme to the “delivery division”scheme, which sorted mail items in the delivery divisions corresponding to each individual postal delivery person. The sorting machine NAS-100, provided in 1997, realized the “route construction division”scheme, which sorted mail items in the delivery route sequence for individual delivery person, by reading seven-digit postal codes and detailed address information together. This scheme was followed to support by TT-200 mail sorting machine in 1998.

For general-purpose applications, Hitachi brought out the first domestic general-purpose OCR product, the H-8252 in 1968. The Electro-Technical Laboratory (at present the National Institute of Advanced Industrial Science and Technology) and Toshiba co-developed the ASPET/71, a document OCR device for English type documents in November 1971. This time frame was the foundation period to decide the OCR business application market. The H-8959 optical character reader, released in June 1972, was the general-purpose OCR device that could read handwritten numbers for the first time in Japan. The Nippon Telegraph and Telephone Public Corporation (at present NTT Data) released a counting-register film reader in October 1973, which could read characters written on the developing film. Other milestones during this period of time included the Mitsubishi Electric M2481 optical character reader in June 1976, which read characters freely handwritten in field boxes on page-length forms (reading multiple lines on one time), and the HITAC T-550/30 OCR system, announced in August 1977, which implemented the distribution system over the network. Fujitsu’s FACOM 6312B, released in August 1978, was the world’s first OCR device to use microprocessors for its control functions, pioneering the future software implementation trends for OCR functions.

The time frame of 1980s was the spread period of OCR business application selection, which gradually expanded and penetrated their target market. The Nippon Telegraph and Telephone Public Corporation (at presence NTT Data) provided the DT-OCR100CN1 model, used in the social insurance data communication system in February 1980, and in the Ministry of Labor system in July 1981. Both systems were spread out nationwide and were well know used to the citizens. As OCR usage expanded widely, large-scale installation models moved toward large-volume high-speed centralized processing systems. On the other hand, distributed models moved toward the small desk-top systems or peripheral devices of general-purpose workstation terminals/PCs, starting with the announcement of the HITAC T-550/47, the first domestic desktop model in March 1984 . Products for special features were also rolled out. The Nippon Telegraph and Telephone Public Corporation (at presence NTT Data) exhibited a handwritten kanji character OCR reader at the International Science and Technology Exposition held in Tsukuba in 1985. This reader was called OCR60 model and integrated into the electronic information processing system to support vehicle registration and inspection services, controlled and operated nationwide by the Ministry of Transport since 1985. This was the first example of the large-scale system application to read handwritten kanji characters. The FACOM6365 was announced in May 1988 to support Japanese language printed documents with flexible formats. This model eliminated any previous restrictions of business form OCR which recognized only the predefined field documents, and realized document OCR which recognized the documents without any consciousness of OCR, such as in newspapers or printed materials.

With handwritten kanji character recognition and document OCR widely available, 1990s were the developing period of OCR business application market selection, which expanded furthermore. Fujitsu developed the color processing OCR of F6335A model in April 1991, which used three RGB light sources. NEC realized the high speed simultaneous reading technology with double-sided OCR in July 1995, which read both front and rear sides of large business forms. In August 1998, NEC brought out the NS-1000, a desk stand type OCR reader, implementing contactless overhead scanner equipment. This product created new business applications such as bank counter customer services. With the wide use of Windows OS PCs and handy flat bed type of scanners, further-inexpensive OCR software products appeared by extracting OCR hardware functions into software implementations. OCR manufacturers for example Fujitsu and Oki Electric switched into the OCR software products, and the specialized software vendors like Media Drive entered into the market. As of the end of 2014, only the three manufacturers — Toshiba, Hitachi, and NEC — provided OCR hardware products.