WO2015189817A2 - Ancient administrative handwritten documents: virtual x-ray reading - Google Patents

Ancient administrative handwritten documents: virtual x-ray reading Download PDF

Info

Publication number
WO2015189817A2
WO2015189817A2 PCT/IB2015/054462 IB2015054462W WO2015189817A2 WO 2015189817 A2 WO2015189817 A2 WO 2015189817A2 IB 2015054462 W IB2015054462 W IB 2015054462W WO 2015189817 A2 WO2015189817 A2 WO 2015189817A2
Authority
WO
WIPO (PCT)
Prior art keywords
ray
projection
page
ink
specimen
Prior art date
Application number
PCT/IB2015/054462
Other languages
French (fr)
Other versions
WO2015189817A3 (en
Inventor
Fauzia ALBERTIN
Giorgio Margaritondo
Frédéric Kaplan
Original Assignee
Ecole Polytechnique Federale De Lausanne (Epfl)
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Ecole Polytechnique Federale De Lausanne (Epfl) filed Critical Ecole Polytechnique Federale De Lausanne (Epfl)
Publication of WO2015189817A2 publication Critical patent/WO2015189817A2/en
Publication of WO2015189817A3 publication Critical patent/WO2015189817A3/en

Links

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/04—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material
    • G01N23/046—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and forming images of the material using tomography, e.g. computed tomography [CT]
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N33/00—Investigating or analysing materials by specific methods not covered by groups G01N1/00 - G01N31/00
    • G01N33/26—Oils; Viscous liquids; Paints; Inks
    • G01N33/32—Paints; Inks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/07—Investigating materials by wave or particle radiation secondary emission
    • G01N2223/076—X-ray fluorescence
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/20—Sources of radiation
    • G01N2223/204—Sources of radiation source created from radiated target
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/30—Accessories, mechanical or electrical features
    • G01N2223/301—Accessories, mechanical or electrical features portable apparatus
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N2223/00—Investigating materials by wave or particle radiation
    • G01N2223/40—Imaging
    • G01N2223/419—Imaging computed tomograph
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/06—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption
    • G01N23/083—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption the radiation being X-rays
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N23/00—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00
    • G01N23/02—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material
    • G01N23/06—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption
    • G01N23/10—Investigating or analysing materials by the use of wave or particle radiation, e.g. X-rays or neutrons, not covered by groups G01N3/00 – G01N17/00, G01N21/00 or G01N22/00 by transmitting the radiation through the material and measuring the absorption the material being confined in a container, e.g. in a luggage X-ray scanners

Definitions

  • the present invention relates to an instrument and the corresponding technology, for the digitization of ancient handwritten documents by x-ray tomography.
  • the invention provides a method for detecting ink writings in a specimen comprising stacked pages, allowing a page-by-page reading without turning pages.
  • the method comprises steps of taking a set of projection x- ray images for different positions of the specimen with respect to an x-ray source and a detector from an apparatus for taking projection x-ray images; storing the set of projection x-ray images in a suitable computer system; and processing the set of projection x-ray images to tomographically reconstruct the shape of the specimen.
  • the method for detecting ink writings further comprises analysing in situ a chemical ink composition of a few characters of the speciment to verify a presence of heavy metals.
  • the method for detecting ink writings further includes the realization of the chemical analysis in situ with the same apparatus used for taking the projection x-ray images.
  • the method for detecting ink writing further comprises steps of applying specialized computer processing for the tomographical reconstruction of the object; and to identify individual pages also in the case of deformations and detect individual characters, words and extended text with the objective of deciphering the content.
  • the method for detecting ink writing further comprises extracting from the tomographically reconstructed shape of the specimen, information on a support and ink morphology suitable for any one of the following list: dating, extracting data for conservation procedures, for paleographic analysis.
  • the taking of projection x-ray images is done exploiting different x-ray contrast mechanisms, including at least one from the list comprising attenuation contrast and different kinds of phase contrast.
  • the invention provides a page-reading setup for reading ink writings in a specimen comprising stacked pages, on a page-by-page basis without turning the pages in the stacked pages, the setup comprising at least a high-flux microfocus x-ray source with an operating voltage below 150 kV; a goniometer configured for handling an accurate positioning and rotation of the specimens; and an imaging device, a two-dimensional x-ray detector based either on CCD or hybrid technologies, with a lateral resolution of at least 300 micron and a size of at least 15x15 cm.
  • the page-reading setup further comprises a mobile support configured to easily transport the setup without disassembling it.
  • the invention enables a new ad hoc x-ray tomographic instrument to investigate, digitize and virtually read un-openable manuscripts.
  • the tomographic data set may be acquired with a minimal handling of the objects and without opening them.
  • a series of x-ray projection images is acquired and computer processed to reconstruct the shape of the object and of the pages. Once the pages are recognized, thanks to the high absorption ink used for ancient handwritings, is possible to recognize the text, and consequently read it.
  • the instrument is configured to be transportable in situ.
  • the inventive technique can read unopened— perhaps unopenable— documents and speed up the entire digitization process of large collections. This is feasible thanks to the highly absorbing chemical elements in ancient ink recipes.
  • the corresponding x-ray contrast allows character reading with reasonably short exposure times. In most cases, the image quality is good enough for tomography reconstruction suitable for virtual page-by-page "reading". When attenuation is too low, phase contrast imaging can reveal the characters from refractive index effects.
  • figure 1 contains a graph showing iron content detected in our seventeen specimens (bottom-side labels) from the intensity of the related x-ray fluorescence (XRF) spectral peaks;
  • figure 2 shows our results for a 1664 Italian (Tuscany) specimen (image size 1 8x 14 mm 2 ); specifically, the bottom-right picture is an x-ray image and the other pictures are standard photographs;
  • figure 3 shows results similar to those of figure 2, for a 15 th century religious writing parchment, of Italian origin (image size 1 8x7 mm 2 ); here, the two bottom pictures are x-ray images;
  • figure 4 shows tomography results for a 1679 Italian (Tuscany) document; the three left-side pictures are standard photographs;
  • figure 5 contains x-ray images taken with different contrast techniques [4,5], for the one of the characters of figure 3—the top picture corresponds to scattering contrast and the bottom to refraction contrast;
  • figure 6 shows results for two sheets of a 1 9 th century legal document.
  • Left side and top-right images visible photographs of the first sheet (recto and verso) and of the second one (recto).
  • Bottom right absorption-contrast x-ray image of the two sheets, taken with a laboratory-based instrument;
  • figure 7 shows results similar to those of figure 6, obtained for a 1 590 private communication handwriting from Italiany.
  • the x-ray image (right) puts well in evidence the paper structure, which can be used as a fingerprint of the paper origin;
  • figure 8 shows a visible photograph (left) of a 1679 accounting manuscript and an x-ray absorption-contrast image taken again with a laboratory-based x-ray source; figure 9 shows the results of a test of "virtual reading” by x-ray tomography; figure 10 shows a test of "virtual reading” by x-ray tomography.
  • the specimen is a 19 th century handwritten book, 400 pages and 5 cm thickness.
  • On the top left a scheme of the tomographic acquisition instrument; the five pictures are examples of the extracted "pages"; and
  • figure 1 1 is a scheme of the instrument treated by the present invention.
  • the instrument includes a microfocus x-ray source, a sample holder with a goniometer for accurate rotations and specimen positioning, a two-dimensional x- ray detection system and a computer system for image processing.
  • VTM Venice Time Machine
  • the project also explores new ways to virtually "read" manuscripts, rapidly and non-invasively.
  • the raw data can be obtained without opening or otherwise manipulating the manuscripts, drastically reducing the risk of damage and speeding up the process.
  • VTM project with the huge size of the corresponding manuscript collection, demonstrates therefore the need for a standardized, general-purpose technique based on x-rays. This goes well beyond the specialized precursor tests of Refs. 1 -3 and requires the specialized instrumentation targeted by the present invention.
  • Clarifying these issues has a remarkable and multi-faceted potential impact.
  • the x-ray attenuation image quality is good enough to perform tomographic reconstruction, as we demonstrated for example for phantom "volumes" created by stacking ancient manuscript fragments.
  • iron gall a name suggesting the presence of iron.
  • This generic formula corresponded to a wide variety of ingredients and recipes [9].
  • the basic fabrication process was the reaction of an acid with an iron compound.
  • the most common procedure involved tannic acid (C76H52O46) and iron sulfate (FeSO- in rainwater, white wine or vinegar.
  • Tannic acid was obtained from plants, the richest source being the “galls” produced by trees in response to parasites attacks (e.g., by gall wasps)— for example, the British oak galls or the top-quality “Aleppo galls”. Iron sulfate was "green vitriol" from mines. The reaction of tannic acid with FeSO4 produced, with oxygen exposure, ferrotannate— a black pigment.
  • iron gall inks are their time evolution. Although a small quantity of black pigment is developed with the oxygen present in the solution, most is produced with atmospheric oxygen after writing, over hours or days. Furthermore, the ink is corrosive over very long periods of time, chemically attacking the substrate.
  • Figure 1 shows iron content detected by x-ray fluorescence (XRF) on seventeen Italian document. We acquired them using the same geometric conditions for all specimens, with a portable ⁇ -XRF spectrometer ARTAX (model 200, Brucker) under He flux, with 200 micron wide spots, operating at 15 kV and 1500 ⁇ (exposure time 180 s).
  • XRF x-ray fluorescence
  • Figures 2 and 3 include a comparison between XRF spectra and x-ray images of a 17 th century Italian writings of personal nature and of a 15 th century religious parchment manuscript.
  • Fig. 2 this shows results for a 1664 Italian (Tuscany) specimen (image size 1 8x 14 mm 2 ).
  • At the top right are shown x-ray fluorescence spectra taken from the 200-micron wide spots marked on the top-left visible picture.
  • Two x-ray effects are visible: attenuation allows character recognition whereas phase contrast enhances the paper fibers.
  • the radiograph reveals otherwise nearly invisible holes caused by ink-induced paper corrosion.
  • Fig. 3 shows results similar to those of Fig. 2, for a 15th century religious writing on parchment, of Italian origin (image size 1 8x7 mm 2 ).
  • the bottom part shows a comparison of two visible pictures of red (left part of Fig. 3) and black (right part of Fig. 3) ink characters (colors not represented in color) and two x- ray absorption images represented respectively below the red and black ink characters.
  • the different amounts of Fe correlate well with the x-ray attenuation contrast, as shown in Figs. 2 and 3.
  • the x-ray obtained images reflect both phase contrast and attenuation contrast.
  • Phase contrast is primarily visible in the microscopic substrate features like paper fibers.
  • Attenuation contrast prevails in the ink areas.
  • FIG. 4 Positive tests of tomographic reconstruction are shown in Fig. 4 for a 1679 specimen, consisting of a stack of eight 0.8 cm diameter fragments simulating a small volume.
  • a 1679 specimen consisting of a stack of eight 0.8 cm diameter fragments simulating a small volume.
  • Figure 4 shows three tomographically reconstructed "pages” with recognizable characters and two three-dimensional reconstructed side views. Note that the spatial resolution (6.5 m detector pixel) is largely sufficient not only to distinguish different "pages” but also to determine if the writing is in the front or in the back of each "page”.
  • contrast mechanisms related to the imaginary part of the complex refractive index i.e., to phase-related phenomena like refraction [7, 1 1 ].
  • Figure 5 shows a test: the imaged area is the same as that of the red characters of Fig. 3 (color not shown in Fig. 3 but characters represented on lower left part of Fig. 3), but the contrast mechanism is different.
  • Figure 3 shows indeed absorption-contrast images whereas Fig. 5 shows instead scattering and refraction images.
  • Figure 6 shows an example of positive result: the x-ray image of two sheets from a 19 th century Italian manuscript.
  • the figure 6 also shows the recto and verso pictures of the top sheet and the recto picture of the bottom sheet (that had no writing in its verso).
  • Characters and words are readily visible in the radiology picture over an approximately 20x10 cm 2 area, and suitable for computerized recognition. We see that characters are visible not only from the recto of the top sheet but also from its verso and from the next sheet. This is due both to the low x-ray absorption of the paper (that primarily contains low-weight chemical elements) and to the strong absorption by the iron contained in the black ink.
  • the superposition of words can be disentangled with x-ray tomography and ad hoc software.
  • the x-ray images carry additional information like the micromorphology of the paper. This aspect is evident from the results of Fig. 7 obtained on a 1590 Italiany manuscript (a private communication). The difference between the visible photograph and the x-ray image is quite striking as far as the paper structure is concerned. Here too, we see in the radiograph characters and words both from the recto and the verso of the paper sheet.
  • Figure 8 shows similar results on a 1679 accounting text, also from Sicily. This was a single sheet not written on the verso. This result shows that radiology with a conventional source can be applied to large areas (here 21 x 15 cm 2 ). The ink contrast produces in the radiograph perfectly readable characters and words. The varying intensity of the ink-related features reveals the ink quantity or density created by the handwriting process— for example, to the re-supplying of the pen or to how the character was written. This is an interesting piece of information for paleography studies.
  • Figure 9 shows two examples of the 360 two-dimensional raw x-ray images taken at equally spaced angles.
  • Figure 9 (middle) shows two tomographic rendered images of the entire stack, with different orientations. From the side of the object, we can observe the thick ink characters.
  • Figure 9 (bottom) presents two examples of tomographic reconstruction of individual inside pages. The object size was approximately 7x5 cm 2 and the thickness of each page was ⁇ 1 50 pm. The entire tomographic set was acquired in «90 minutes. Higher current levels will decrease the total acquisition time.
  • Figure 1 0 shows a scheme of part of the instrument including the book holder and the tomographic reconstructed images of some pages. We took the tomographic set in a total time of ⁇ 60 minutes.
  • the inventive method for detecting ink writings may be applied to a large variety of specimens, including all types of written or printed specimens consisting of stacked pages: handwritten or printed books or portions of books, folded documents, sealed documents, scrolls, pages embebbed in other items or otherwise conceiled from direct view.

Landscapes

  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Chemical & Material Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Biochemistry (AREA)
  • Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
  • Pulmonology (AREA)
  • Radiology & Medical Imaging (AREA)
  • Theoretical Computer Science (AREA)
  • Food Science & Technology (AREA)
  • Medicinal Chemistry (AREA)
  • Apparatus For Radiation Diagnosis (AREA)
  • Analysing Materials By The Use Of Radiation (AREA)

Abstract

A method for detecting ink writings in a specimen comprising stacked pages, allowing a page-by-page reading without turning pages The method compris- es steps of taking a set of projection x-ray images for different positions of the specimen with respect to an x-ray source and a detector from an apparatus for taking projection x-ray images; storing the set of projection x-ray images in a suitable computer system; and processing the set of projection x-ray images to tomographically reconstruct the shape of the specimen.

Description

Ancient Administrative Handwritten Documents: Virtual X-ray Reading Technical field
The present invention relates to an instrument and the corresponding technology, for the digitization of ancient handwritten documents by x-ray tomography.
Background
In reading ancient handwritings, there are precursors projects exploiting x- rays to decipher documents. Such projects targeted only specific items without developing a general-purpose technology suitable for many different kinds of specimens and for large collections. Notably, synchrotron light was used to retrieve "lost" text from the "Archimedes Palimpsest" [1 ] with x-ray fluorescence. Wess targeted instead "unrolled" scrolls producing flat readable images and assessing radiation damage.[2,3].
Summary of the invention
In a first aspect the invention provides a method for detecting ink writings in a specimen comprising stacked pages, allowing a page-by-page reading without turning pages. The method comprises steps of taking a set of projection x- ray images for different positions of the specimen with respect to an x-ray source and a detector from an apparatus for taking projection x-ray images; storing the set of projection x-ray images in a suitable computer system; and processing the set of projection x-ray images to tomographically reconstruct the shape of the specimen.
In a preferred embodiment, the method for detecting ink writings further comprises analysing in situ a chemical ink composition of a few characters of the speciment to verify a presence of heavy metals.
In an even further preferred embodiment, the method for detecting ink writings further includes the realization of the chemical analysis in situ with the same apparatus used for taking the projection x-ray images.
In an even further preferred embodiment, the method for detecting ink writing further comprises steps of applying specialized computer processing for the tomographical reconstruction of the object; and to identify individual pages also in the case of deformations and detect individual characters, words and extended text with the objective of deciphering the content. In an even further preferred embodiment, the method for detecting ink writing further comprises extracting from the tomographically reconstructed shape of the specimen, information on a support and ink morphology suitable for any one of the following list: dating, extracting data for conservation procedures, for paleographic analysis.
In an even further preferred embodiment, the taking of projection x-ray images is done exploiting different x-ray contrast mechanisms, including at least one from the list comprising attenuation contrast and different kinds of phase contrast.
In a second aspect, the invention provides a page-reading setup for reading ink writings in a specimen comprising stacked pages, on a page-by-page basis without turning the pages in the stacked pages, the setup comprising at least a high-flux microfocus x-ray source with an operating voltage below 150 kV; a goniometer configured for handling an accurate positioning and rotation of the specimens; and an imaging device, a two-dimensional x-ray detector based either on CCD or hybrid technologies, with a lateral resolution of at least 300 micron and a size of at least 15x15 cm.
In an even further preferred embodiment, the page-reading setup further comprises a mobile support configured to easily transport the setup without disassembling it.
Hence the invention enables a new ad hoc x-ray tomographic instrument to investigate, digitize and virtually read un-openable manuscripts.
The tomographic data set may be acquired with a minimal handling of the objects and without opening them.
A series of x-ray projection images is acquired and computer processed to reconstruct the shape of the object and of the pages. Once the pages are recognized, thanks to the high absorption ink used for ancient handwritings, is possible to recognize the text, and consequently read it.
Beside the tomographic investigation, a parallel study of the chemical characteristics of the inks is done using the same instrument.
Given to the nature of the object analyzed and the scope of the invention, ancient handwritings - often fragile - and large collection, the instrument is configured to be transportable in situ.
The inventive technique can read unopened— perhaps unopenable— documents and speed up the entire digitization process of large collections. This is feasible thanks to the highly absorbing chemical elements in ancient ink recipes. The corresponding x-ray contrast allows character reading with reasonably short exposure times. In most cases, the image quality is good enough for tomography reconstruction suitable for virtual page-by-page "reading". When attenuation is too low, phase contrast imaging can reveal the characters from refractive index effects.
The potential impact is quite revolutionary: the feasibility of "softer" and faster digitization of huge collections. One example is the "Archivio di Stato" in Venice, objective of our Venice Time Machine project. However, our invention is generally applicable to any collection of manuscripts with sufficient x-ray imaging contrast, such as the vast majority of European specimens from the 1 3th to 20th centuries.
Brief description of the drawings
The invention will be better understood in view of the description of preferred example embodiments and in reference to the figures, wherein
figure 1 contains a graph showing iron content detected in our seventeen specimens (bottom-side labels) from the intensity of the related x-ray fluorescence (XRF) spectral peaks;
figure 2 shows our results for a 1664 Italian (Tuscany) specimen (image size 1 8x 14 mm2); specifically, the bottom-right picture is an x-ray image and the other pictures are standard photographs;
figure 3 shows results similar to those of figure 2, for a 15th century religious writing parchment, of Italian origin (image size 1 8x7 mm2); here, the two bottom pictures are x-ray images;
figure 4 shows tomography results for a 1679 Italian (Tuscany) document; the three left-side pictures are standard photographs;
figure 5 contains x-ray images taken with different contrast techniques [4,5], for the one of the characters of figure 3— the top picture corresponds to scattering contrast and the bottom to refraction contrast;
figure 6 shows results for two sheets of a 1 9th century legal document. Left side and top-right images: visible photographs of the first sheet (recto and verso) and of the second one (recto). Bottom right: absorption-contrast x-ray image of the two sheets, taken with a laboratory-based instrument;
figure 7 shows results similar to those of figure 6, obtained for a 1 590 private communication handwriting from Tuscany. The x-ray image (right) puts well in evidence the paper structure, which can be used as a fingerprint of the paper origin;
figure 8 shows a visible photograph (left) of a 1679 accounting manuscript and an x-ray absorption-contrast image taken again with a laboratory-based x-ray source; figure 9 shows the results of a test of "virtual reading" by x-ray tomography; figure 10 shows a test of "virtual reading" by x-ray tomography. The specimen is a 19th century handwritten book, 400 pages and 5 cm thickness. On the top left, a scheme of the tomographic acquisition instrument; the five pictures are examples of the extracted "pages"; and
figure 1 1 is a scheme of the instrument treated by the present invention. The instrument includes a microfocus x-ray source, a sample holder with a goniometer for accurate rotations and specimen positioning, a two-dimensional x- ray detection system and a computer system for image processing.
Detailed description of preferred example embodiments
I. Reading Ancient Handwritings: Role of X-rays
In order to establish the foundations of our approach, we performed extensive studies to correlate the chemical analysis of 15th -20th century administrative Italian documents and the x-ray imaging of the handwritten characters and words. Heavy elements in the inks [6] normally allow their detection by x-ray attenuation. But strong concentration fluctuations occur between manuscripts and even within the same specimen. Sometimes, the low concentration impedes detection by attenuation: an alternate method is refractive-index contrast [7].
Our study is related to the international Venice Time Machine (VTM) project [8]. The Archivio di Stato in Venice holds about 80 kms of archival documents spanning over ten centuries and documenting every aspect of the Venetian Mediterranean Empire. If unlocked and transformed in a digital data system, this information could change significantly our understanding of European history. But the sheer mass of data is a problem. VTM plans to digitalize and decipher the entire collection in 10-20 years.
To facilitate and accelerate this task, the project also explores new ways to virtually "read" manuscripts, rapidly and non-invasively. We specifically plan to use x-ray tomography to computer-extract page-by-page information from sets of projection images. The raw data can be obtained without opening or otherwise manipulating the manuscripts, drastically reducing the risk of damage and speeding up the process.
The VTM project, with the huge size of the corresponding manuscript collection, demonstrates therefore the need for a standardized, general-purpose technique based on x-rays. This goes well beyond the specialized precursor tests of Refs. 1 -3 and requires the specialized instrumentation targeted by the present invention.
Along the path to deciphering the Venice collection, we must deal with a series of key issues. A fundamental one is the nature of the ink in everyday documents (as opposed to pieces of high artistic or historical value) and its detection by x-rays. This is part of the present description.
The issue is by no means clear a priori. The very nature of the VTM project - and of the analysis of European collections in general - requires studying archival documents such as ship records, notary papers, work contracts, tax declarations, commercial transactions or demographic accounts. For such items, the ink composition is scarcely documented - or not documented at all. Furthermore, the Archivio di Stato collection spans ten centuries, with inevitable fluctuations in the ink chemistry.
Clarifying these issues has a remarkable and multi-faceted potential impact. First, in addition to the Venice collection the techniques could be applied to many documents at risk throughout the world. Second, the chemical and mi- crostructural information also clarifies deterioration mechanisms and could help preventing them.
Our main results are the following. First, heavy elements are systematically detected in the inks of personal and commercial documents over the six centuries we investigated. Second, the chemical composition is basically consistent with the historical records about the inks [9]. Third, the relative amounts of heavy elements change drastically between different document areas and from manuscript to manuscript, with no historical trend. Fourth, there is— as expected— a direct correlation between the quality of attenuation- contrast x-ray images and the chemical composition.
In many cases, the x-ray attenuation image quality is good enough to perform tomographic reconstruction, as we demonstrated for example for phantom "volumes" created by stacking ancient manuscript fragments.
Our initials tests were performed with centralized synchrotron radiation sources of x-rays. However, the objective of a general-purpose technique is in evident conflict with the use of centralized facilities. The use of non- centralized sources is indeed a crucial question since moving specimens to a centralized facility creates conservation, logistic, practical and even legal problems that would severely reduce the applicability of the approach. One crucial step in our feasibility tests was, therefore, the demonstration that our invention can be implemented using conventional x-ray sources. This was done, as shown for example by Figs. 6-10. Such figures present radiography and tomography results successfully obtained with a laboratory- based x-ray source on single-page manuscripts and tomography of a modern sample simulating a 10-page volume. Most significantly, we present good tomographic results on a 19th handwritten book, with a thickness of 5 centimeter and some 400 pages.
Our feasibility tests also demonstrated that the technique can be applied large-areas specimens. This is another positive result of avoiding the use of synchrotron sources, since synchrotron emission is by its very nature extremely collimated and typically limited to small specimens.
II. Details on the Chemistry of Ancient Inks and Its Role in X-ray Tomography
Why were heavy elements present in ancient inks? Let us start from black inks. For many centuries, Europe used a formula generically denominated as "iron gall", a name suggesting the presence of iron. This generic formula, however, corresponded to a wide variety of ingredients and recipes [9]. The basic fabrication process was the reaction of an acid with an iron compound. The most common procedure involved tannic acid (C76H52O46) and iron sulfate (FeSO- in rainwater, white wine or vinegar.
Tannic acid was obtained from plants, the richest source being the "galls" produced by trees in response to parasites attacks (e.g., by gall wasps)— for example, the British oak galls or the top-quality "Aleppo galls". Iron sulfate was "green vitriol" from mines. The reaction of tannic acid with FeSO4 produced, with oxygen exposure, ferrotannate— a black pigment.
One important property of iron gall inks is their time evolution. Although a small quantity of black pigment is developed with the oxygen present in the solution, most is produced with atmospheric oxygen after writing, over hours or days. Furthermore, the ink is corrosive over very long periods of time, chemically attacking the substrate.
The above features agree with our chemical analysis.
Figure 1 shows iron content detected by x-ray fluorescence (XRF) on seventeen Italian document. We acquired them using the same geometric conditions for all specimens, with a portable μ-XRF spectrometer ARTAX (model 200, Brucker) under He flux, with 200 micron wide spots, operating at 15 kV and 1500 μΑ (exposure time 180 s).
Figures 2 and 3 include a comparison between XRF spectra and x-ray images of a 17th century Italian writings of personal nature and of a 15th century religious parchment manuscript. Referring to Fig. 2, this shows results for a 1664 Italian (Tuscany) specimen (image size 1 8x 14 mm2). At the top right are shown x-ray fluorescence spectra taken from the 200-micron wide spots marked on the top-left visible picture. At the bottom is shown a comparison of a 14x 1 8 mm2 visible picture (left) with a synchrotron x-ray image. Two x-ray effects are visible: attenuation allows character recognition whereas phase contrast enhances the paper fibers. The radiograph reveals otherwise nearly invisible holes caused by ink-induced paper corrosion.
Fig. 3 shows results similar to those of Fig. 2, for a 15th century religious writing on parchment, of Italian origin (image size 1 8x7 mm2). The bottom part shows a comparison of two visible pictures of red (left part of Fig. 3) and black (right part of Fig. 3) ink characters (colors not represented in color) and two x- ray absorption images represented respectively below the red and black ink characters.
All spectra revealed Fe in the ink areas (except the 1881 ), as opposed to the substrate areas. In addition, we detected Ca in the same ink areas. This is consistent with the use of gum Arabic as the binder. Calcium was also found in the substrate of the parchment specimens presumably due to chalk used during support preparation [10].
These rather straightforward results are accompanied by several less obvious findings. In the 1664 (Fig. 2) and in the parchment specimens (Fig. 3), the ink also contains other elements (like K, Cu and Zn), attributed to vitriol impurities. Furthermore, the Fe and Ca contents varied substantially.
The different amounts of Fe correlate well with the x-ray attenuation contrast, as shown in Figs. 2 and 3. The x-ray obtained images reflect both phase contrast and attenuation contrast. Phase contrast is primarily visible in the microscopic substrate features like paper fibers. Attenuation contrast prevails in the ink areas.
In addition to black (iron gall) ink, ancient manuscripts could also have color inks. Our 15th century parchment specimen included red characters, written with ink containing Hg (cinnabar). Hg was indeed detected in the x-ray fluorescence spectra, as seen in Fig. 3, and produced excellent attenuation contrast. On the contrary, the Fe signal from the black characters is weak, and the contrast low.
Positive tests of tomographic reconstruction are shown in Fig. 4 for a 1679 specimen, consisting of a stack of eight 0.8 cm diameter fragments simulating a small volume. We took a set of 1 ,001 projection images at equally spaced angles over a π radian range, using 15 keV photons and a time exposure of 10 ms per picture.
Figure 4 shows three tomographically reconstructed "pages" with recognizable characters and two three-dimensional reconstructed side views. Note that the spatial resolution (6.5 m detector pixel) is largely sufficient not only to distinguish different "pages" but also to determine if the writing is in the front or in the back of each "page".
III. Alternate Imaging Modes
How could we handle the cases of low attenuation contrast? We could use contrast mechanisms related to the imaginary part of the complex refractive index— i.e., to phase-related phenomena like refraction [7, 1 1 ].
Figure 5 shows a test: the imaged area is the same as that of the red characters of Fig. 3 (color not shown in Fig. 3 but characters represented on lower left part of Fig. 3), but the contrast mechanism is different. Figure 3 shows indeed absorption-contrast images whereas Fig. 5 shows instead scattering and refraction images.
The important point here is that the refraction images reflect the local specimen morphology rather than only its chemical composition. We can conclude that what we see in Fig. 5 reflects the substrate morphology modified by the writing process or by the ink-substrate interaction. This could be used to detect a character when the attenuation contrast would give a faint picture.
IV. Radiography of large area manuscripts
As already argued, tests based on synchrotron light could not be directly applied for a massive digitization project. We therefore tested the technique with a conventional x-ray source, allowing to perform large-area investigation and in situ analysis.
Figure 6 shows an example of positive result: the x-ray image of two sheets from a 19th century Tuscany legal manuscript. For comparison, the figure 6 also shows the recto and verso pictures of the top sheet and the recto picture of the bottom sheet (that had no writing in its verso). Characters and words are readily visible in the radiology picture over an approximately 20x10 cm2 area, and suitable for computerized recognition. We see that characters are visible not only from the recto of the top sheet but also from its verso and from the next sheet. This is due both to the low x-ray absorption of the paper (that primarily contains low-weight chemical elements) and to the strong absorption by the iron contained in the black ink. The superposition of words can be disentangled with x-ray tomography and ad hoc software.
The x-ray images, including those using conventional sources, carry additional information like the micromorphology of the paper. This aspect is evident from the results of Fig. 7 obtained on a 1590 Tuscany manuscript (a private communication). The difference between the visible photograph and the x-ray image is quite striking as far as the paper structure is concerned. Here too, we see in the radiograph characters and words both from the recto and the verso of the paper sheet.
Figure 8 shows similar results on a 1679 accounting text, also from Tuscany. This was a single sheet not written on the verso. This result shows that radiology with a conventional source can be applied to large areas (here 21 x 15 cm2). The ink contrast produces in the radiograph perfectly readable characters and words. The varying intensity of the ink-related features reveals the ink quantity or density created by the handwriting process— for example, to the re-supplying of the pen or to how the character was written. This is an interesting piece of information for paleography studies.
The entire test procedure required minimal handling of the manuscripts. Throughout it, we found no evidence whatsoever of x-ray-induced damage, to the supports or to the inks. This is consistent with our knowledge of x-ray physics and chemistry.
V. Tomography of large area manuscripts
For this test, we used a simulated specimen consisting of a stack of ten pages written with iron-gall ink on parchment-like paper. Figure 9 (top) shows two examples of the 360 two-dimensional raw x-ray images taken at equally spaced angles. Figure 9 (middle) shows two tomographic rendered images of the entire stack, with different orientations. From the side of the object, we can observe the thick ink characters. Figure 9 (bottom) presents two examples of tomographic reconstruction of individual inside pages. The object size was approximately 7x5 cm2 and the thickness of each page was ~ 1 50 pm. The entire tomographic set was acquired in «90 minutes. Higher current levels will decrease the total acquisition time.
VI. Tomography of a 400 pages book
We also performed tomography on a 400 page book using again a conventional x-ray source.
Figure 1 0 shows a scheme of part of the instrument including the book holder and the tomographic reconstructed images of some pages. We took the tomographic set in a total time of ~ 60 minutes.
From Fig. 1 0 we see that, even without an optimized software for automatic computer extraction of pages, it is possible to look inside the volume and see writings, letter and words in each of the 400 pages. VII. Conclusions and Perspectives
The above tests carry positive messages about our invention. Rather simple procedures already yielded good results for a variety of manuscript-related problems. The only obstacle was the impossibility to image inks without heavy metals by absorption-contrast radiology and tomography. On the contrary, tests on colored inks containing other heavy metals such as mercury were even more successful than those on iron gall inks.
In some instances the heavy element content of the iron gall ink was too weak. Our preliminary tests indicate that these cases can be potentially handled with phase-contrast images [7, 1 1 ].
Our x-ray images carry a wealth of information besides the written characters. They notably reveal details of the ink-substrate interaction, which in many cases leads to corrosion and deterioration. This will conceivably identify ways to prevent long-term damage.
Results as those discussed here open the way to a new strategy for information harvesting from ancient documents— alternate to page-by-page taking of visible pictures. A single tomographic set could yield the same information more rapidly and with minimized interaction with the document. The computer storage of information in this form can be handled with the present technologies.
Throughout our tests we found no evidence whatsoever of radiation damage, consistent with Ref. 3.
This opens the door to wide applications, which will be based on ad hoc instruments covered by the present application (Fig. 1 1 ), suitable for use at the manuscript collection sites and optimized for this task— rather than the conversion of systems conceived for other applications. The use of ad hoc instruments is, in particular, essential for the commercial use of the techniques as a service to third parties.
The inventive method for detecting ink writings may be applied to a large variety of specimens, including all types of written or printed specimens consisting of stacked pages: handwritten or printed books or portions of books, folded documents, sealed documents, scrolls, pages embebbed in other items or otherwise conceiled from direct view. References
1 . www.slac.stanford.edu/gen/com/images/technical%20summary_final.pdf
2. Mills, D., Samko, 0., Rosin, P., Thomas, K., Wess, T., Davis, G. R.
(2012) . Developments in X-ray Tomography VIII, Proc. SPIE 8506, 85060A.
3. Patten, K., Gonzales, L, Kennedy, C, Mills, D., Davis, G. R., Wess, T.,
(2013) . Heritage Science 1 , 1 -9.
4. Weitkamp, T., Diaz, A., David, C, Pfeiffer, F., Stampanoni, M., Cloetens, P., Ziegler, E. (2005). Optics Express 13, 6296-6304.
5. Stampanoni, M., Groso, A., Isenegger, A., Mikuljan, G., Chen, Q., Meis- ter, D., Lange, M., Betemps, R., Henein, S., Abela, R. (2007). Synch. Rad. Instr, Pts land 2, 879, pp 848 - 851
6. Del Carmine, P., Giuntini, L, Hooper, W., Lucarelli, F., Mando, P. A.
(1996). Nucl. Instrum. Methods B113, 354-356.
7. Hwu, Y., Hsieh, H. H., Lu, M. J., Tsai, W. L, Lin, H. M., Goh, W. C, Lai, B., Je, J. H., Kim, C. K., Noh, D. Y., Youn, H. S., Tromba, G., Margari- tondo, G. (1999). J. Appl. Phys. 86, 4613-4618.
8. http://dhvenice.eu/
9. Library Special Collections Conservator Unit, Preservation Department, Yale University, Medieval Manuscripts, Some Ink and Pigment Recipes (http://travelingscriptorium.files.wordpress.com/2012/03/scopa-recipes- booklet_web.pdf) (2012).
10. Van der Snickt, G., De Nolf, W., Vekemans, B., & Janssens, K. (2008).
Applied Physics A, 92(1 ), 59-68.
1 1 . Margaritondo, G., Hwu, Y. (2013). Nuovo Saggiatore 29, 45-58.

Claims

Claims
1 . Method for detecting ink writings in a specimen comprising stacked pages, allowing a page-by-page reading without turning pages, comprising steps of
taking a set of projection x-ray images for different positions of the specimen with respect to an x-ray source and a detector from an apparatus for taking projection x-ray images; storing the set of projection x-ray images in a suitable computer system; and
processing the set of projection x-ray images to tomographically reconstruct the shape of the specimen.
2. The method for detecting ink writings of claim 1 , further comprising analysing in situ a chemical ink composition of a few characters of the speciment to verify a presence of heavy metals.
3. The method for detecting ink writings of claim 2, further including the realization of the chemical analysis in situ with the same apparatus used for taking the projection x-ray images.
4. The method for detecting ink writing of claim 1 , further comprising steps of
applying specialized computer processing for the tomographical reconstruction of the object; and
to identify individual pages also in the case of deformations and detect individual characters, words and extended text with the objective of deciphering the content.
5. The method for detecting ink writing of claim 1 , further comprising extracting from the tomographically reconstructed shape of the specimen, information on a support and ink morphology suitable for any one of the following list: dating, extracting data for conservation procedures, for paleographic analysis.
6. The method for detecting ink writings of claim 1 , wherein the taking of projection x-ray images is done exploiting different x-ray contrast mechanisms, including at least one from the list comprising attenuation contrast and different kinds of phase contrast [7, 1 1 ].
A page-reading setup for reading ink writings in a specimen comprising stacked pages, on a page-by-page basis without turning the pages in the stacked pages, the setup comprising at least
a high-flux microfocus x-ray source with an operating voltage below 150 kV;
a goniometer configured for handling an accurate positioning and rotation of the specimens;
an imaging device, a two-dimensional x-ray detector based either on CCD or hybrid technologies, with a lateral resolution of at least 300 micron and a size of at least 15x15 cm.
The page-reading setup of claim 4, further comprising a mobile support configured to easily transport the setup without disassembling it.
PCT/IB2015/054462 2014-06-13 2015-06-12 Ancient administrative handwritten documents: virtual x-ray reading WO2015189817A2 (en)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
IBPCT/IB2014/062219 2014-06-13
IB2014062219 2014-06-13

Publications (2)

Publication Number Publication Date
WO2015189817A2 true WO2015189817A2 (en) 2015-12-17
WO2015189817A3 WO2015189817A3 (en) 2016-03-03

Family

ID=53761434

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/IB2015/054462 WO2015189817A2 (en) 2014-06-13 2015-06-12 Ancient administrative handwritten documents: virtual x-ray reading

Country Status (1)

Country Link
WO (1) WO2015189817A2 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110634099A (en) * 2018-06-25 2019-12-31 夏普株式会社 Book electronization device and book electronization method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9068920B2 (en) * 2009-01-14 2015-06-30 John Eric Churilla System and method for scanning and processing printed media

Non-Patent Citations (8)

* Cited by examiner, † Cited by third party
Title
"Library Special Collections Conservator Unit, Preservation Department, Yale University, Medieval Manuscripts, Some Ink and Pigment Recipes", 2012
DEL CARMINE, P; GIUNTINI, L.; HOOPER, W.; LUCARELLI, F.; MANDO, P. A, NUCL. INSTRUM. METHODS, vol. B113, 1996, pages 354 - 356
HWU, Y.; HSIEH, H. H.; LU, M. J.; TSAI, W. L.; LIN, H. M.; GOH, W. C.; LAI, B.; JE, J. H.; KIM, C. K.; NOH, D. Y., J. APPL. PHYS., vol. 86, 1999, pages 4613 - 4618
MARGARITONDO, G.; HWU, Y, NUOVO SAGGIATORE, vol. 29, 2013, pages 45 - 58
PATTEN, K.; GONZALES, L.; KENNEDY, C.; MILLS, D.; DAVIS, G. R.; WESS, T., HERITAGE SCIENCE, vol. 1, 2013, pages 1 - 9
STAMPANONI, M.; GROSO, A.; ISENEGGER, A.; MIKULJAN, G.; CHEN, Q.; MEIS-TER, D.; LANGE, M.; BETEMPS, R.; HENEIN, S.; ABELA, R., SYNCH. RAD. INSTR, PTS LAND, vol. 2, no. 879, 2007, pages 848 - 851
VAN DER SNICKT, G.; DE NOLF, W.; VEKEMANS, B.; JANSSENS, K., APPLIED PHYSICS A, vol. 92, no. 1, 2008, pages 59 - 68
WEITKAMP, T.; DIAZ, A.; DAVID, C.; PFEIFFER, F.; STAMPANONI, M.; CLOETENS, P.; ZIEGLER, E., OPTICS EXPRESS, vol. 13, 2005, pages 6296 - 6304

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110634099A (en) * 2018-06-25 2019-12-31 夏普株式会社 Book electronization device and book electronization method

Also Published As

Publication number Publication date
WO2015189817A3 (en) 2016-03-03

Similar Documents

Publication Publication Date Title
Lak et al. Phase contrast X-ray synchrotron imaging: opening access to fossil inclusions in opaque amber
Albertin et al. X‐ray spectrometry and imaging for ancient administrative handwritten documents
Albertin et al. Ancient administrative handwritten documents: X-ray analysis and imaging
Mazzeo Analytical chemistry for cultural heritage
Gueriau et al. Trace elemental imaging of rare earth elements discriminates tissues at microscale in flat fossils
Christiansen et al. The nature of ancient Egyptian copper-containing carbon inks is revealed by synchrotron radiation based X-ray microscopy
Limburg et al. Insights from two‐dimensional mapping of otolith chemistry
Wilster‐Hansen et al. Virtual unwrapping of the BISPEGATA amulet, a multiple folded medieval lead amulet, by using neutron tomography
WO2015189817A2 (en) Ancient administrative handwritten documents: virtual x-ray reading
VIEGAS et al. PRELIMINARY STUDIES ON IRON GALL INKS COMPOSITION USING AN EXTERNAL ION BEAM.
Müller et al. Proton beam irradiation induces invisible modifications under the surface of painted parchment
Manfriani et al. The forger's identikit: A multi-technique characterization of Pippo Oriani's fake paintings
Pandey et al. Forensic Investigation of Suspected Document for Alteration, Erasures & Obliteration
Cohen et al. Composition analysis of writing materials in Geniza fragments
Lee et al. TOF-SIMS study of red sealing-inks on paper and its forensic applications
Albertin X-ray tomography for manuscripts
Albertin et al. From closed testaments to books: Virtual X-ray Reading as an alternate digitization technology for fragile documents
Li et al. An examination of handwritten signatures forged using photosensitive signature stamp
Bergmann et al. Chemical mapping of ancient artifacts and fossils with x-ray spectroscopy
Milota et al. PIXE measurements of Renaissance silverpoint drawings at VERA
Bergmann et al. Chemical mapping of ancient artifacts and fossils with X-ray spectroscopy
Edwards et al. Photons, folios, and fossils: The X-ray Imaging and Spectroscopy Program of Ancient Materials at SSRL
Dowsett et al. XEOM 1-A novel microscopy system for the chemical imaging of heritage metal surfaces
Wainwright Examination of paintings by physical and chemical methods
Oliaiy et al. Analysis of Iranian postage stamps belonging to the Qajar dynasty (18th–20th century's) by micro‐PIXE

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 15744343

Country of ref document: EP

Kind code of ref document: A2

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 15744343

Country of ref document: EP

Kind code of ref document: A2