Historical document image database is developed in collaboration with the National Archives of Tunisia and contains 10 000 images which have been scanned in grayscale mode (8 bits / pixel) and in color mode (24 bits / pixel), at a resolution of 300 dpi which is acceptable for the most of ancient documents, with variable sizes using the "TIFF" and "JPEG" formats. This database contains Tunisian historical documents which date from the Ottoman period (1574-1881), the colonial period (1881-1956) and the independent Tunisia (since 1956), and represent manuscripts, printed documents and periodical documents in Arabic, French, Italian and English languages. Currently, our database images are manually annotated using 970 keywords spread over about fifteen annotation classes with typically 15-30 keywords per image. These keywords are associated with each image using an appropriate XML file allowing a better description and an easier representation of the various images.
Samples of the SID HDI database:
Téléchargement: license agreement SID HDI Database