Introduction
What I am going to tell you happened in Brittany on the 9th of July 2004, when we were on holiday. It was after dinner and I had just put a piece of aluminum foil on a plate of cheese to cover it. I know … It’s very prosaic and my mind was far from any experiment but this is what happened.
As I was sitting at the table chatting with my mother and my husband—I still can’t believe it—I looked at the recently covered cheese plate and my daughter’s face was printed on that foil just in front of us. I then started to take photos of this very amazing event which seemed truly incredible to me. While I was taking the photographs with my Minolta Dimage XT, my mother became nervous and thought it was silly that I photographed a plate and asked me why I was doing it. I got up and went near her to show her my small camera screen without telling her anything. She immediately recognized the face of her granddaughter and asked me where I had got it. I told her it was on the aluminium paper; she then got up and came to where I was, sat on my chair and without me telling her the place she immediately spotted this little face. Then the same thing happened with my husband. And my husband—who is not a good observer—exclaimed: “It’s incredible; I can see Bénédicte’s face!” and he burst into tears.
I took the photos because I was afraid to see the image disappear. But the truth is that the little face remained on the same spot of the aluminium foil until the 14th of July and in the meantime I took photos with another camera, a Nikon F-401. The Nikon was less convenient to photograph the little face because it does not zoom and I wanted to show the photos to other people….
On the 14th of July we had to return because our holidays were over and I had to hand back the plate which was not mine, it belonged to the house we were renting and so I unwrapped the aluminum foil. To this day I regret not having brought that plate with me for I think I could still have my daughter’s face imprinted on the aluminum foil! It was really silly of me not having brought the plate home. Many people, friends who saw the photos later on, had the same opinion: they could see Bénédicte’s face on them.
Marie-Hélène Bienaimé, France
Physiognomic Analysis Performed on an
Image Obtained through I. P. V. S.*
by Daniele Gullà”Interdisciplinary Laboratory for Biopsychocybernetics Research
* I.P.V.S. : Acronym devised by IL Laboratorio which stands for “Interazione Psi-Visiva di tipo Strumentale” (Psi-Visual Interaction of Instrumental type, more commonly known as ITC image).
The Case Under Investigation
The investigation consists of attempts to explore the authenticity of an alleged ITC image unexpectedly obtained when a plate covered with aluminum foil was photographed. In the photograph, the creases in the foil were apparently seen to have spontaneously taken on the visual configuration of the face of a deceased girl. The girl’s identity was immediately claimed to be recognized by her mother (who took the photograph) and by other relatives present at the time.
Description of the Case
The anomalous photograph that is at the center of the case initially came to the attention of Dr. Anabela Cardoso, who then arranged for it to be sent to me for analysis. It was shown to Dr. Cardoso by the lady responsible for taking it and whom Dr. Cardoso and I met while attending the International Conference of Infinitude held in Paris in 2007. Having inspected the photograph, Dr. Cardoso considered it of sufficient interest to be sent to me, together with photographs taken during the lifetime of the deceased daughter whose image it appeared to be. These photographs would enable me to carry out a possible verification of any similarities in physiognomic characteristics between the alleged ITC image on the tinfoil and the images of the girl shown in the photographs taken during her lifetime.
Technical Steps in Making the Anthropometric Comparison
The ITC image and one of the lifetime pictures of Bénédicte were normalized in terms of pixels and contrast, and subsequently, metric and morphologic measurements were taken of the two images. The two images were then superimposed one on the other so that the appropriate mark points (‘repère’ points) coincided. Morphologic Test of Pattern Recognition with Neural Networks: The test of morphologic compatibility was performed with the Neural Networks program utilized by the American intelligence services (FBI) known as the Universal Image Recognition program. This morphologic test takes into consideration the whole of the cranial structure of the human head, elaborating the contours of the “structural cage” and does not focus on the distances of single repère points. Instead it analyzes in a more general way the shape of the various parts of the cranium. The test was performed using the ONE_TO_MANY mode that compares the ITC image with 3,000 other images contained in the database of the Universal Image Recognition program. This database is composed of masculine and feminine somatic types of faces, all of European origin and aged between 10 and 70 years.
The final data is the result of a miscellaneous comparison of 3,000 x 3,000 or more precisely 9,000,000 comparative tests. In the final result visible on the screen in Figure 7, seven similar images were found but only one reaches the maximum score which surpasses the threshold of FAR (False Acceptance) and FRR (False Recognition Rate), which the program sets respectively at percentage values 0, 1% and 0, 03%. This image was one of Bénédicte’s lifetime photographs. That image reached the highest score and it was therefore identified as the face most similar to the ITC image with a rate of 98.97%.
 |
 |
 |
| Figure 2. Detail of the supposed ITC face enlarged and rotated through approximately 30 degrees | Figure 3. Four images of Bénédicte, the deceased French girl visually recognized by her mother as the subject of the ITC image in the aluminum foil, taken during her lifetime | Figure 4. This image shows the result of the superimposition of the lifetime image of Bénédicte’s face with the presumed ITC face. The similarity of the somatic traits is clearly noticeable. |
It is possible to read on the central report of the computer screen printout in Figure 7 that besides the choice of the anomalous ITC image proper (which was also added to the database of the program), which is identical to the image for which a comparison was required because it is the same image, and therefore attained the highest score of 192000000 (file denominated “Photo David et Bénédicte1r.jpg”), the second most similar image with a score of 59136000 is Bénédicte’s lifetime photograph (file denominated “Photo David et Bénédicte-2.jpg”) shown on the right side of the computer print-out. This image, which gave a percentage similarity rating of 98.7% was therefore identified as the face most similar to the ITC image (The Italian Courts of Justice accept a reading of 95% when establishing cases of human identity).
 |
 |
 |
| Figure 5. The same operation was then carried out using another image of Bénédicte superimposed on the ITC picture with alignment of the repère points with morphing technique. The similarity of the faces and the apparent coincidence of the points are again noticeable. | Figure 6. In this analysis the distance relationships between the repère points of the two images were measured. Although the relationships are noticeably constant, with values in the region of number 1, there are percentage value changes that indicate the presence of some spatial deformations in the tinfoil of the ITC image that prevent definitive conclusions, from the metric point of view, as to whether or not the two images share the same identity. | Figure 7. Morphologic comparison between the ITC image and 3,000 faces performed with neural networks. Bénédicte’s lifetime image was recognized as the most similar to the anomalous image with a percentage of 98.97%. |
Conclusions
The technological applications used in the comparative analysis have revealed several points of similarity. The morphologic comparison done with Neural Networks on a sampling of 3,000 faces shows that a high percentage of compatibility between the ITC image and the photograph taken during the young girl’s life exists. However, it should be borne in mind that
the alleged ITC image taken by the mother under the circumstances and with the exposure value (i.e. the time the shutter remained open during the shutter click) concerned, produced a particular effect of light/shadow on the reflective surface (i.e. the aluminum foil) that does not allow us to highlight identification marks capable of allowing an evaluation sufficiently sensitive to yield total compatibility with an image taken during the subject’s lifetime. Furthermore, it should be emphasized that my analysis is limited to measurements made on optical information whose authenticity and origin cannot be accurately determined.
Unfortunately, it was not possible to achieve an accurate comparison of the metric relationship between the points because, although a very marked similarity was found, the image containing the “extra” presents spatial deformations which are obviously due to the particular conditions of the creased aluminum foil on which the image was imprinted. Consequently, it was not possible to obtain precise anthropometric data from the image. Proper measurements could only be taken from Bénédicte’s lifetime pictures. The missing instrumental confirmation of the metric methodology does not, however, diminish the probability of a definite identification, even though we have to deal with comparisons between two images that differ from each other as to the material on which they are imposed. It would have been different if we had analyzed two homogeneous photographs and had found instrumental discrepancies with one of the two methods used.
Nevertheless, the morphologic analysis (which is more representative by virtue of its analysis of the images through the simulation of human vision) and the probabilistic comparison between the ITC image and the population of 3,000 faces in the database render the hypothesis of compatibility of the two faces valid.
Translated from the Italian original by Dr. Anabela Cardoso and originally published in No. 31 of the ITC Journal which is published 3 times a year. For subscription information, see itcjournal.org
Bibliography
[1] Bertillon A., Identification anthropometrique: instruction signaletiques, Melun, Imprimerie Administrative, 1893.
[2] Goldstein A., Harmon L.D., Lesk A.B., Identification of Human faces, Proceeding of the IEEE, Vol, 59, No.5, MAY 1971.
[3] Duda R., Hart P., Stork G., Pattern classification, Wiley, New York, 2000.
[4] Watanabe S., Pattern recognition:human and mechanical, Wiley, New York, 1985.
[5] Falco G., Identità: metodo scientifico di segnalamento e identificazione, II Edizione, Roma, 1923.
[6] Martin R., Saller K., Lehrbuch der Anthropologie in systematicher Darstellung, Dd. I-II. Fischer, Stuttgard, 1956-59.
[7] Reverte-Coma G.M., Antropologia forense, Ministero de Justitia, Madrid, 1991.
[8] Schiwidwtzky I., Knussmann R., Morphognose und Typognose, in Martin R. Antropologie, Fischer, Stuttgard, 1988.
[9] Iscan M.Y., Loth S.R., Photo image identification, in Siegel J.A., Soukko J.P., Knupfer G.C., (Eds), Encyclopedia of Forensic Sciences, Academic Press, pagg 795-805, New York, 2000.
[10] Olivieri L., Antropologia e Antropometria, C.E.V. Idelson, Napoli, 1963.
[11] Farkas L. G., Anthropometry of the head and face, Second Edition, Raven Press, 2000.
[12] Parisi, R., Nozioni di antropometria, EDAS, Messina, 1979.
[13] Balossino N., Siracusa S., Parametri discriminatori nel riconoscimento di volti, Polizia Moderna, n.1, 1998.
[14] Chellapa R., Wilson C.L., Sirohey S., Human and machine recognition of faces: a survey, Proceedings of the IEEE, 83(5):705-735, may 1995.
[15] Brunelli R., Poggio T., Face recognition:features versus Templates, IEEE Transaction on Pattern analysis and machine intelligence,vol. 15, no. 10, october 1993.
[16] Wu C., Huang J., Human face profile recognition by computer, Pattern recognition, Vol 23 No.3/4 pp.255-259, 1990.
[17] Wong K.H., Hudson H.M., Law and Tsang P.W.M., A system for recognising human faces, Dept. of Computer Science, The Chinese University of Hong Kong and Dept. of Elect. Eng. City Poly. of U. K., Technical Report, 1991.
[18] I. Craw, P. Cameron. Face recognition by computer. Pagg 498-508. D.Hogg and R. Boyle editors, Proceedings British Machine Vision Conference, Springer Verlag, 1992.
[19] M.S. Kamel, H.C. Shen, A.K.C. Wong, T.M. Hong, R.I. Campenu, Face recognition using perspective invariant features, Pattern Recognition Letters, 877-884, 1994.
[20] P. Rajesh, N. Rao, H. Ballard, Natural basis functions and topographic memory for face recognition, Proceedings of the International Conference of Artificial Intelligence, 10-17, 1995.
[21] D. Gullà, “Two Cases of an Anthropometric Identification in Two Supposed Anomalous Photographs”, Proceedings of Congress the Association “Casa dell’ Albero”, October 2001, Carpi (MO)
[22] D. Gullà, “Forensic Analyses in Parapsichology“, Proceedings of XVI National Congress of “Il Movimento della Speranza”, January 2002, Roma.
[23] D. Gullà, “Images and Voice Anomalous; Case of Pattern Recognition,” session of University of Imola (BO), Gen. 2004
