Steven Avery
Administrator
217-232 with scientific machinery and pictures of ms.
===========================================
P. 233-246 with machinery AND PICTURES - ENGLISH
This essay presents efforts taken to analyse the manuscript with the shelf mark Suppl. gr. 119 housed in the Austrian National Library. The manuscript is a palimpsest, whereby the underwritten text runs horizontally and the overwritten text runs vertically. While it is known that the underwritten text is a forgery, created by Constantine Simonides, it is not yet known whether the overwritten text is also a forgery or a genuine historical text.
.............
While the works mentioned above are concerned with contrast enhancement,
Luo et al.9 have recently presented a method for forgery detection in hyperspectral
document images. Contrary to the other methods described, their
method has been applied to modern documents for which ground truth data
exists. The overall aim of the method is to automatically distinguish between
visually similar inks. For this purpose three different anomaly detection
methods are compared and it is shown that they are capable of correctly estimating
the number of ink types and to distinguish between the different writings.
The inks contained in the examples look similar in the visible range, but
have a different spectral signature. The approach is especially designed for such
types of documents.
Contrary, in our specimen Suppl. gr. 119 the horizontally and vertically
written layers have a different appearance in images taken in the visible range.
The overall question in this essay is whether the entire vertical text is forged and
not whether it was written with different inks. Thus, the work in Luo et al.10 is
related to our work because it deals with forgery detection, but not directly
applicable to the manuscript investigated in this paper.
..............
Conclusion
This essay dealt with the analysis of the palimpsest manuscript Suppl. gr. 119 of
the Vienna National Library. While it is known that the palimpsest text in this
manuscript is a forgery, the question remains whether the overwritten text is
also a forgery. Since multispectral imaging has proven its usefulness for the
analysis of historical manuscripts, the particular object has been analysed with
this non-invasive investigation technique. By making use of MSI and several
post-processing techniques the contrast of the forged palimpsest text was enhanced,
compared to regular white light illumination. As a result, the over- and
the underwriting showed different spectral signatures. Hence it was possible to
separate them and to visualise the difference between the two inks by generating
pseudo colour images. Neither in the multispectral images nor in the corresponding
spectral curves any conspicuous details were detected. Consequently,
based on the techniques described here, no statement about the origin of the
overwriting in ms. Suppl. gr. 119 can be made; although it should be added that
the extreme regularity of the two layers (their degree of visibility in enhanced
images, the rectangular relationship of the layers, etc.) differs from the genuine
historical evidence we have observed so far.
===========================================
===========================================
P. 233-246 with machinery AND PICTURES - ENGLISH
This essay presents efforts taken to analyse the manuscript with the shelf mark Suppl. gr. 119 housed in the Austrian National Library. The manuscript is a palimpsest, whereby the underwritten text runs horizontally and the overwritten text runs vertically. While it is known that the underwritten text is a forgery, created by Constantine Simonides, it is not yet known whether the overwritten text is also a forgery or a genuine historical text.
.............
While the works mentioned above are concerned with contrast enhancement,
Luo et al.9 have recently presented a method for forgery detection in hyperspectral
document images. Contrary to the other methods described, their
method has been applied to modern documents for which ground truth data
exists. The overall aim of the method is to automatically distinguish between
visually similar inks. For this purpose three different anomaly detection
methods are compared and it is shown that they are capable of correctly estimating
the number of ink types and to distinguish between the different writings.
The inks contained in the examples look similar in the visible range, but
have a different spectral signature. The approach is especially designed for such
types of documents.
Contrary, in our specimen Suppl. gr. 119 the horizontally and vertically
written layers have a different appearance in images taken in the visible range.
The overall question in this essay is whether the entire vertical text is forged and
not whether it was written with different inks. Thus, the work in Luo et al.10 is
related to our work because it deals with forgery detection, but not directly
applicable to the manuscript investigated in this paper.
..............
Conclusion
This essay dealt with the analysis of the palimpsest manuscript Suppl. gr. 119 of
the Vienna National Library. While it is known that the palimpsest text in this
manuscript is a forgery, the question remains whether the overwritten text is
also a forgery. Since multispectral imaging has proven its usefulness for the
analysis of historical manuscripts, the particular object has been analysed with
this non-invasive investigation technique. By making use of MSI and several
post-processing techniques the contrast of the forged palimpsest text was enhanced,
compared to regular white light illumination. As a result, the over- and
the underwriting showed different spectral signatures. Hence it was possible to
separate them and to visualise the difference between the two inks by generating
pseudo colour images. Neither in the multispectral images nor in the corresponding
spectral curves any conspicuous details were detected. Consequently,
based on the techniques described here, no statement about the origin of the
overwriting in ms. Suppl. gr. 119 can be made; although it should be added that
the extreme regularity of the two layers (their degree of visibility in enhanced
images, the rectangular relationship of the layers, etc.) differs from the genuine
historical evidence we have observed so far.
===========================================
Last edited: