From the NDL:
On March 29, 2019, the NDL launched its “Next Digital Library” content-based retrieval system prototype that has advanced functions using artificial intelligence (AI). The NDL is developing this system so future digital library services will be more easily usable.
“Next Generation Digital Library” is an experimental search service that implements the functions developed based on research in the Next Generation System Development Laboratory of the National Diet Library.
Among the materials provided by the National Diet Library Digital Collection, some of the materials (10,550 points) that have a copyright protection period expired and which are classified into the Japanese decimal classification (NDC) 6 class (industry) It is searchable. The materials to be searched will be added one by one.
It has the following features:
Search by Keyword allows you to search for material from the full text generated by the OCR. The hit part is displayed in the form of a snippet.
“Find by Pictures” allows you to search for images and illustrations that have been automatically extracted from the document. By selecting a picture, you can search for materials that contain similar images.
In the viewer of the document, it is possible to display in page unit divided automatically by the throat line, in addition to the usual spread display.
It can be downloaded after performing image correction (whitening) to remove noise etc. during scanning. Processing time (about 15 seconds per image) is required. In addition, corrections for photos and figures may not work because of the function mainly aimed at improving the readability of the text part.
No English-language search interface is available. However, it’s possible to get an idea of how the search tool works and looks by using English-language keywords and clicking “around” the results page. Here’s a search for the word “music”.