From closed testaments to books: Virtual X-ray Reading as an alternate digitization technology for fragile documents

Abstract

In recent years, research and technology made considerable progress in increasing the speed and the safety of the entire digitization process of ancient collections. Despite this, imaging ancient, fragile or un-opened documents remains a formidable challenge. We employ an alternate digitization technique for handwritten documents, exploiting x-ray tomography: Virtual X-ray Reading. Thanks to the high penetration of x-rays, we can acquire 3-dimensional (3D) tomographic images of a multi-page document without opening it. The x-ray contrast necessary for the readability is produced by the chemical composition and the consequent strong x-ray absorption of the iron gall inks -largely used for European handwritten documents. We present the development of this technology, from the chemical investigations of the inks to the tomography of an unopened Venetian testament and of an 18 th century, 200-page, handwritten book.

Unlocking history through automated virtual unfolding of sealed documents imaged by X-ray microtomography

“…Our virtual unfolding pipeline builds on prior work using XMT to extract text hidden inside sealed, damaged, or otherwise unopenable historical documents 18 . Recent investigations lay out a computational framework for virtually unwrapping or unrolling these documents to a flat isometry so their contents can be more readily observed; to date these investigations have focused primarily on scrolls [10][11][12][19][20][21][22][23][24][25][26][27] , books 13,28,29 , and artifacts folded once or twice 28,30,31 . The primary motivation for many prior studies has been to recover text from highly damaged documents that cannot be physically opened and read.…”

From tomographic reconstruction to automatic text recognition: the next frontier task for the artificial intelligence

“…These works are a prime example of alignment of paper documents using an algorithm based on camera calibration, regardless of the image content. The image of the text fragments has been successfully restored for parchment scrolls [16][17][18], scrolled and folded papyri [19][20][21][22][23][24][25], a silver scroll [26,27] and a folded lead amulet [28,29]. Much of these works are devoted to methods and algorithms for unfolding scrolls in one piece, which have a non-symmetric structure.…”

“…Howe ver, in subsequent works, the process was fully automated and allowed not only to unfold the scrolls and discover the text on them, but also to improve the visualization of the material, which makes studying easier than ever before. Also the text fragments has been successfully restored for books and booklets [23,[30][31][32][33], several times folded paper documents and sealed letters [34]. The ability to access hidden text fragments has been checked for model objects in the form a scroll of bamboo planks [35] and a scrolled piece of birch bark [36].…”

A Geometric Feature-Based Algorithm for the Virtual Reading of Closed Historical Manuscripts

“…In this direction, several techniques have been proposed, from a fully manual approach [29], to a combination of algorithmic and user-driven operations [30], to a more automatic technique exploiting segmentation, surface modeling and ink projection [31], or the use of a triangular mesh for surfaces characterization [23,32]. More automatic algorithms, such as the topological identification and propagation, were applied by [27,33] on scrolls and, more recently, by [25,26] on book and bamboo scrolls.…”