Next Digital Library (English)
The Next Digital Library is an experimental search service that implements functions developed based on research conducted by the Research and Development for Next-Generation Systems Office in the National Diet Library. Its purpose is to verify the technical effectiveness of the full-text search function, automatic processing using machine learning, International Image Interoperability Framework (IIIF) API, and more.
The "Keyword search" allows users to search from the full text, and the "Illustration search" allows users to search for similar images and illustrations automatically extracted from each material.
Digitalized materials of books including old and rare Japanese books (about 336,000 items) with expired copyright protection available on the internet in The National Diet Library Digital Collection can be searched.
Operation checks are performed with the latest Chrome and Firefox. It does not work with Internet Explore.
The functions offered are as follows: 1. Full-text search, 2. Image search using automatic image extraction, 3. Background whitening of offered materials, 4. Automatic generation of table of contents in materials, 5. Automatic image processing for smartphone display, 6. Automatic detection of page turning direction, 7. Utilization of IIIF, 8. Automatic cropping of document image margins in the browser, and 9. Adding tags related to the content using image recognition and faceted search by tags.
For more details, please see the documents listed on NDL Lab related literature (Japanese).
Also, related programs and learning datasets used for machine learning are released on GitHub (https://github.com/ndl-lab).