Text Readability and Coreference Annotation across Heterogeneous Media for the Digital Archive of Rare Books

This paper focuses on the public digital archive of rare books. The main issue is the improvement of utility in terms of text information, which is the most valuable contents of rare books. We argue that balance between fidelity and utility is an important notion to deal with the problem of reproducibility, which is problematic only in public digital archives. We propose an image processing method called ``digital bleaching'' to improve text utility in terms of text readability without taking the risk of reproducibility. Next we discuss that the generation of electronic text using optical character recognition leads to the improvement of text utility in terms of text searchability. Thirdly, we propose a new software called Open Ontology Forge (OOF), which is designed to make annotations on rich semantic relationship called ``coreference'' across heterogeneous media such as between text and image, which is common in rare books with pictures and drawings. Finally, we have already started the construction of the digital archive based on the achievement of our research.

Citation

Asanobu KITAMOTO, Takeo YAMAMOTO, Sonoko SATO, Nigel COLLIER, Ai KAWAZOE, Kinji ONO, "Text Readability and Coreference Annotation across Heterogeneous Media for the Digital Archive of Rare Books", The Journal of the Institute of Image Electronics Engineers of Japan, Vol. 33, No. 5, pp. 737-745, 2004-09 (in Japanese)

BibTeX Format

@Article{ kyscko:iieej04,
	author = {Asanobu KITAMOTO and Takeo YAMAMOTO and Sonoko SATO and Nigel COLLIER and Ai KAWAZOE and Kinji ONO},
	title = {Text Readability and Coreference Annotation across Heterogeneous Media for the Digital Archive of Rare Books},
	journal = {The Journal of the Institute of Image Electronics Engineers of Japan},
	volume = {33},
	number = {5},
	pages = {737-745},
	year = 2004,
	month = 09,
	note = { (in Japanese)},
}

Related Resources and Related Websites

Related Pages in this Site

| Link 1 |