ATSI Bari, 5 September 2014
Shroud-like coloration, conservation
Shroudmeasures and image processing
A survey of experiments at ENEA Frascati
Paolo Di Lazzaro
Lazzaro, Daniele Murra
ENEA Center of Frascati
paolo dilazzaro@enea it
paolo.dilazzaro@enea.it
Outline
¾ The
Th images
i
on the
th Shroud,
Sh
d STuRP
ST RP
¾ Photochemistry-based linen coloration
¾ Conservation measures
¾ Misleading image processing
¾ Conclusion
Paolo Di Lazzaro, ATSI 2014 Bari
The images on the Shroud
Recent Photo of the Turin Shroud (Courtesy David Rolfe)
The two faint bodyy images
g have a lower contrast and are less
visible than all the stains thereabouts: blood, water, burns.
Paolo Di Lazzaro, ATSI 2014 Bari
The STuRP Project
Torino, 8‐14 October, 1978.
32 scientists and technicians, 2.5 million $ equipment, 120 hours no‐stop analyses,
2 years for data elaboration
Paolo Di Lazzaro, ATSI 2014 Bari
Main findings of STuRP
The Shroud is not a painting, no pigment, any directionality, not a scorch.
g encodes cloth to body
y distance,, and it is p
present in both contact
The image
and non contact areas.
The image is superficial, no more than 0.6 microns thick (work by others
)
has shown 0.2 microns).
Invisible halos surround blood.
Blood went on before
image
(no
image
beneath blood).
The blood stains contain
hemoglobin and serum
albumin.
Calcium and strontium
and iron are uniformly
present on the Shroud
p
in small quantities.
Paolo Di Lazzaro, ATSI 2014 Bari
STuRP conclusion
“We
We can conclude for now that the
Shroud image is that of a real
human form of a scourged,
crucified man.
man
It is not the product of an artist.
The blood stains are composed of
hemoglobin and also give a positive
test for serum albumin.
The image is an ongoing mystery
and until further chemical studies
are made, perhaps by this group of
scientists or perhaps by some
scientists,
scientists in the future, the
problem remains unsolved.”
Paolo Di Lazzaro, ATSI 2014 Bari
Outline
¾ The
Th images
i
on the
th Shroud,
Sh
d STuRP
ST RP
¾ Photochemistry-based linen coloration
¾ Conservation measures
¾ Misleading image processing
¾ Conclusion
Paolo Di Lazzaro, ATSI 2014 Bari
Why photochemistry? Why UV?
Energy carried
E
i d by
b short-wavelength
h
l
h radiation
di i
b k chemical
breaks
h i l
bonds of the irradiated material without inducing a significant
heating (photochemical reaction).
reaction)
Moreover, linen has a molar absorptivity which increases when
decreasing the radiation wavelength: the smaller the wavelength,
the thinner the material necessary to absorb all the radiation.
Then, we have
Th
h
chosen
h
th ultraviolet
the
lt i l t radiation
di ti as an “acting
“ ti att
distance” mechanism to obtain at least two of the main
characteristics of the Shroud image: a thin coloration depth
and a low-temperature image-formation.
Paolo Di Lazzaro, ATSI 2014 Bari
How much different is our linen from the Shroud?
The solid lines show the absolute reflectance of the linen of the Shroud in areas of
no-image as a function of the wavelength (Gilbert, Appl. Opt. 1980).
The dashed
Th
d h d line
li shows
h
the
h absolute
b l reflectance
fl
off the
h linen
li
used
d in
i our experiments.
i
From J. Imag. Sci. Techn. 54, 4302 (2010)
Paolo Di Lazzaro, ATSI 2014 Bari
Set-up at ENEA Frascati
Hercules ENEA
PBUR
6 J, 120 ns, 5 Hz
308 nm
LPX 305 PBUR
LPX-305,
0.5 J, 30 ns, 50 Hz. 308 nm o 193 nm
Paolo Di Lazzaro, ATSI 2014 Bari
Ultrashort UV pulses irradiate linens
Paolo Di Lazzaro, ATSI 2014 Bari
Short movie
Excimer laser irradiation of linen
Paolo Di Lazzaro, ATSI 2014 Bari
Linen thread after 10ns, λ = 193nm
irradiation
From J. Imag. Science Techn. 54, 040201(2010)
Paolo Di Lazzaro, ATSI 2014 Bari
Depth of coloration: UV vs. VUV
λ = 193 nm
Cross section of
irradiated linen
threads
λ = 308 nm
Di Lazzaro, Regina Apostolorum 9/4/14
Paolo Di Lazzaro, ATSI 2014 Bari
Sometimes VUV does not color the scw
inside the fiber
N the
Note
h small
ll yellowish
ll i h pieces
i
off broken
b k pcw
From Applied Optics 51, 8567 (2012)
Paolo Di Lazzaro, ATSI 2014 Bari
Latent images generated by VUV
Unpublished
Paolo Di Lazzaro, ATSI 2014 Bari
Half-tone effect by VUV irradiation
Microscope view of linen threads after VUV laser irradiation. Single
colored fibers are visible next to uncolored fibers
fibers, like in the Shroud
image (areal density coloration).
From Applied Optics 51, 8567 (2012)
Paolo Di Lazzaro, ATSI 2014 Bari
Cold or thermal coloration?
λ = 308 nm
T = 21 °C - 34 °C
λ = 193 nm
T = 22 °C - 25 °C
Paolo Di Lazzaro, ATSI 2014 Bari
Short movie
Th excimer
The
i
laser
l
as a contactless
t tl brush
b h
L
BS
EXCIMER LASER
ℓ
PD
OS
Paolo Di Lazzaro, ATSI 2014 Bari
Shroud-like face using excimer laser as a brush
The face image produced by laser shots is very faint, invisible at sunlight.
Only in shadow we can perceive a low-contrast , yellowish faint image.
image
Unpublished
Paolo Di Lazzaro, ATSI 2014 Bari
Shroud-like face using excimer laser as a brush
The negative of the previous photos shows a positive image.
The negative allows to see the single laser shots producing the very faint image.
Since 2013
Si
it is exposed
i th
in
the Museum
M
of the Shroud
i Turin
in
T i
Unpublished
Di Lazzaro, Regina Apostolorum 9/4/14
Paolo Di Lazzaro, ATSI 2014 Bari
Outline
¾ The
Th images
i
on the
th Shroud,
Sh
d STuRP
ST RP
¾ Photochemistry-based linen coloration
¾ Conservation measures
¾ Misleading image processing
¾ Conclusion
Paolo Di Lazzaro, ATSI 2014 Bari
From cellulose to cromophore
(conjugated carbonyl groups)
aldehyde ‐CHO Alkene –C=C – and ketonic carbonyls – C=O absorb λ << 260 nm
absorb λ
<< 260 nm
absorb λ < 320 nm (large bandwidth)
absorb λ
< 320 nm (large bandwidth)
Aldehyde absorbs both 193nm and 308nm to generate cromophores. But only 193nm is
absorbed by alkene and ketonic groups, thus delocalising groups and shifting absorption
band to longer wavelengths
wavelengths, in the blue-green spectral region
region. This causes the yellow
coloration.
J. Appl. Polymer Science 16, 2567-2576 (1972). Cellulose 1, 205-214 (1994)
Paolo Di Lazzaro, ATSI 2014 Bari
Ionizing radiation vs. Shroud conservation
Gas Radon is the most dangerous source of natural ionizing radiation, potentially
affecting the long term conservation of the image contrast and visibility.
Gases and pressure
Our
suggestions
99.6% Ne or Ar,
0.4% O2,
9 mbar water vapor,
P = 1,05 bar
Present
reliquary
li
99.5% Ar,
0.5% O2
P equalized to the
atmospheric pressure
Relative
humidity,
temperature
RH ≈ 40%
T = 20 °C
RH = 50%
T = 19 °C
Box material /
thickness
Al or Al-based
alloy
t ≥ 5 mm
Al alloy,
t variable
Radon issue
Building
materials and
objects around
must be Radonfree
Information not
available
From EAI, special issue on Knowledge, Diagnostics and Preservation of Cultural Heritage,
pp. 89-94 (2012).
Paolo Di Lazzaro, ATSI 2014 Bari
Outline
¾ The
Th images
i
on the
th Shroud,
Sh
d STuRP
ST RP
¾ Photochemistry-based linen coloration
¾ Conservation measures
¾ Misleading image processing
¾ Conclusion
Paolo Di Lazzaro, ATSI 2014 Bari
A hidden face on the back of the Shroud?
J. Opt. A: Pure Appl. Opt. vol. 6 pp. 491-503
(2004)
Deep image
i
processing
i
off a screenedd
photographic reproduction of the backside
taken from a book. Image processing tools
i l d d convolution
included
l ti
with
ith Gaussian
G
i filters,
filt
summation of images, gamma correction and
filtering in spatial frequency by direct and
inverse 2D Fourier transforms.
transforms
Sindon vol. 19 pp. 57-69 (2003)
Direct and inverse Fourier transform of a highresolution image directly obtained by in
in-depth
depth
scanning of the backside of the Shroud, did not show
any face or any other image.
Paolo Di Lazzaro, ATSI 2014 Bari
A hidden face on the back of the Shroud?
A direct, in‐depth analysis on the Shroud shows there is no hidden face or any other image
on the backside of the Shroud. How can we crosscheck this result?
We checked both images on the left have the same
histogram of “number of pixels” vs. “gray levels”. Then,
after the optimum spatial overlap of the two images
(using bloodstains as a reference), the absolute value of
the difference of gray levels between images should give
a black figure when the two images are perfectly
correlated, or a noisy figure when they have a partial
degree of correlation, or a recognizable face when
their correlation is poor.
PIXEL PERCENTAGE
1.2 ‐
BACK SIDE
1.0 ‐
0.8 ‐
FRONT SIDE
0.6 ‐| | | | | |
0
50 100 150 200 250
0.4 ‐
GREY LEVEL
Paolo Di Lazzaro, ATSI 2014 Bari
A hidden face on the back of the Shroud?
A direct, in‐depth analysis on the Shroud shows there is no hidden face or any other image
on the backside of the Shroud. How can we crosscheck this result?
We checked both images on the left have the same
histogram of “number of pixels” vs. “gray levels”. Then,
after the optimum spatial overlap of the two images
(using bloodstains as a reference), the absolute value of
the difference of gray levels between images should give
a black figure when the two images are perfectly
correlated, or a noisy figure when they have a partial
degree of correlation, or a recognizable face when
their correlation is poor.
The result of this pixel-by-pixel calculation is
shown on the left,
left where a Shroud-like face can be
still recognized, thus revealing the lack of spatial
correlation of the images above, as a whole.
From Pattern Recognition 46, 1964 (2013)
Paolo Di Lazzaro, ATSI 2014 Bari
No hidden face on the back of the Shroud
The lack of spatial correlation between the “hidden” face backside and the face on the
Shroud is further confirmed byy a template
p
matchingg based on a cross-correlation algorithm
g
that calculates the Pearson product-moment correlation coefficient S according to the
following equation:
= 0.4
04
where Aij (Bij) is the gray level of the single pixel (i,j)
(i j) of the image A (B),
(B) and <A> (<B>) is
the mean value of the gray level of the image A (B).
When applying this equation to the front and “hidden” face on the back we obtain S = 0.4
S = 0.4 is a very poor spatial correlation value, typical of completely different images.
As an example,
p , Figures
g
on the left have the
same correlation coefficient S = 0.4 of the
front and presumed back faces on the
Shroud
Paolo Di Lazzaro, ATSI 2014 Bari
Why we perceive a face on the back of the
Shroud?
In summary, both a direct, in-depth analysis on the Shroud and two different methods of
spatial correlation show there is no hidden face on the backside of the Shroud. It remains the
problem to understand why most of us can perceive a Shroud-like
Shroud like face on the back,
back if it
does not exist.
The p
presumed face on the back actuallyy contains 3 ‘faces’ at least…
The face-like patterns we perceive is possibly due to a combined effect of the brain’s ability
to supply the missing contours in order to make sense out of any pattern we can see and of
our innate ppropensity
p
y to interpret
p stimuli as ffaces based on minimal cues. As a matter off
fact, we are hard-wired to perceive faces from several random patterns!
Paolo Di Lazzaro, ATSI 2014 Bari
Our psychophysiology mechanism ‘sees’ faces
everywhere
y
Paolo Di Lazzaro, ATSI 2014 Bari
How it works? Retrieving capability of
our brain
Can you read me?
Can y
you read me?
When a phrase is in a context, our brain is able to retrieve its meaning even when
it is
i half
h lf cancelled
ll d (Gestalt:
(G t lt perception
ti is
i nott a simple
i l addition
dditi off the
th elements
l
t we
see, but it is the result of the relations among the observed objects).
A single letter, when half cancelled, cannot be retrieved (detail vs. context).
C4N Y0U R34D M3?
The same ability of retrieving incomplete information holds for “looking like”
patterns (numbers instead of letters).
Paolo Di Lazzaro, ATSI 2014 Bari
Outline
¾ The
Th images
i
on the
th Shroud,
Sh
d STuRP
ST RP
¾ Photochemistry-based linen coloration
¾ Conservation measures
¾ Misleading image processing
¾ Conclusion
Paolo Di Lazzaro, ATSI 2014 Bari
Conclusion
• We have presented the first comprehensive survey of the
overall research work on Shroud-related issues done in the
last 8 years in the ENEA Research Center of Frascati.
• Our irradiations of linen fabrics show the ability of ultra-short,
high intensity UV and VUV radiation to reproduce many
peculiar aspects of the microscopic complexity of the Turin
Shroud images.
• The different coloration results obtained in different irradiation
conditions allow to recognize photochemical reactions involved
in our Shroud
Shroud-like
like coloration, thus offering hints to plan the
long-term conservation of both cloth and image of the TS.
• Our study of the misleading effects of digital image processing
off low-contrast
l
t
t images
i
confirms
fi
th t there
that
th
i a narrow
is
boundary between image enhancement and manipulation. Our
results corroborates previous works showing the absence of
any face on the backside of the Shroud.
Paolo Di Lazzaro, ATSI 2014 Bari
Contributions from
Giuseppe
pp Baldacchini
Paolo Di Lazzaro
Giulio Fanti
Daniele Murra
Enrico Nichelatti
Antonino Santoni
Barrie Schwortz
ENEA Frascati
ENEA Frascati
P d University
Padua
U i
it
ENEA Frascati
ENEA Casaccia
ENEA Frascati
STERA Inc.
who gave a contribution to different phases of this
long-term work and coauthored papers published in
peer reviewed journals
Paolo Di Lazzaro, ATSI 2014 Bari
Deeper insights may be found in
• G. Baldacchini, P. Di Lazzaro, D. Murra, G. Fanti: “Coloring linens with
excimer lasers to simulate the body image of the Turin Shroud” Applied
Optics 47, 1278-1283 (Optical Society of America, 2008).
• P. Di Lazzaro, D. Murra, A. Santoni, G. Fanti, E. Nichelatti, G. Baldacchini:
“Deep
Deep Ultraviolet radiation simulates the Turin Shroud image
image” Journal of
Imaging Science and Technology 54, 040302-(6) (Imaging.org 2010).
• D. Murra, P. Di Lazzaro: “Sight
g and brain, an introduction to the visually
y
misleading images”, International Workshop on the Scientific approach to
the Acheiropoietos Images, edited by P. Di Lazzaro (ENEA 2010) pp. 31 - 34.
•
P. Di Lazzaro,
P
Lazzaro D.
D Murra,
Murra A.
A Santoni,
Santoni E.
E Nichelatti,
Nichelatti G.
G Baldacchini:
“Superficial and Shroud-like coloration of linen by short laser pulses in
the vacuum ultraviolet” Applied Optics 51, 8567-8578 (Optical Society of
A
America,
i 2012).
2012)
•
P. Di Lazzaro, D. Murra, B. Schwortz: “Pattern recognition after image
processing of low
low-contrast
contrast images, the case of the Shroud of Turin
Turin”
Pattern Recognition 46, 1964-1970 (Elsevier, 2013).
Paolo Di Lazzaro, ATSI 2014 Bari