Google will invest in R & D of text recognition to boost the e-Library

Google said it would sponsor the ai r & D group and will jointly develop text feature recognition technology in the future.
According to CNET, this project is an open resource type named Ocropus and has several main objectives, including "developing an easy-to-use handwriting recognition system to convert handwritten documents into computer text, it helps Google better build its own e-Library, analyze historical documents, and help users with vision problems read information." It is reported that the first three letters of the project name OCR represents the meaning of optimal text recognition.
The Project R & D team is located in the image understanding and pattern recognition Research Group of the German ai r & D center. On a recent Google blog, the search giant launched the above plan. In addition to Google's R & D sponsorship, the group also received investment from multiple German government departments as well as government and private institutions.
The Ocropus team predicts that the project will last for the next three years and will sponsor three doctoral students and post-doctoral students. Prior to this, there were two research achievements in this field, namely the handwriting recognition system proposed by the U.S. Bureau of Population Statistics in the 1990s S and the latest layer analysis method. This research will be based on these two results. On the other hand, other text recognition projects sponsored by Google will also provide resource assistance to the group, for example, Tesseract-one of which was first proposed by HP, in addition, Google improved the open-source optimal text recognition technology last year.
Currently, the preview version of the Ocropus system can be downloaded on the project webpage. The project owner said that, considering the nature of open resources, he hoped to recruit more volunteers to complete the system. For example, you can create a desktop application for the system, add a third-party tool, and use the Ocropus system in multiple languages. Currently, it can only support English.