Penselskriftsanalyse af J.L. Jensens malerier Autenticitet og forfalskninger. Bilag

Transkript

1 Penselskriftsanalyse af J.L. Jensens malerier Autenticitet og forfalskninger Susan Ritterband Konservatorskolen Det Kgl. Danske Kunstakademi 2006

2 Forsidefotografi: J.L. Jensen, Røde liljer og fuchsia, udsnit. Bruun Rasmussens auktion nr. 743, katalognr. 1030

3 Indholdsfortegnelse over bilag 1. Indledning Copydan, jurist Kirsten Kierkegaard, mail Kunsten som narrespeil, hjemmeside Todays Art Forgery, hjemmeside J.L. Jensen Theobald Steins erindringer, Erik Steins afskrift, mail Replikker og kopier af malerier af J.L. Jensen og af hans elever Litteratur om penselskriftsanalyse og tangerende analysemetoder A digital technique for art authentication, artikel The Authentic Project, hjemmeside Bevaringschef, konservator Jørgen Wadum, mail Van Dantzigs liste over Van Goghs karakteristika J.L. Jensens penselskrift udledt af referencemalerier Referencemateriale, J.L. Jensen, Statens Museum for Kunst KMS 220, Tilgang ved det Kgl. Billedgallerie KMS 220, kartotekskort KMS 285, Tilgang ved det Kgl. Billedgallerie KMS 285, kartotekskort KMS 449, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 449, kartotekskort KMS 546, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 546, kartotekskort KMS 1062, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 1062, kartotekskort KMS 1062, sagførers håndskrevne kopi af brev fra van Osten, uddrag Referencemateriale, J.L. Jensen, Thorvaldsens Museum B 230, kartotekskort B 231, kartotekskort B 232, kartotekskort B 230, B 231 og B 232, brev til Thorvaldsen, 1834, B 230, B 231 og B 232, kvittering til Thorvaldsen for køb Transskription af J.L. Jensens brev og kvittering til Thorvaldsen B 230, B 231 og B 232, Thorvaldsens Museum, Haandkatalog for besøgende, B 234, kartotekskort B 234, Thorvaldsens Museum, Haandkatalog for besøgende, Referencemateriale, Kunstakademiets Samling, KS 92, Kunsthistoriker, mag.art. Emma Sallings registrering Mail fra seniorforsker, mag.art. Claus M. Smidt J.L. Jensens penselskrift appliceret på negative kontrolmalerier Hanne Hellesen, Statens Museum for Kunst KMS 305, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 305, kartotekskort KMS 430, Inventarium over Malerierne i Den Kgl. Malerisamling

4 35. KMS 430, kartotekskort Hermania Neergaard, Thorvaldsens Museum B 279, kartotekskort B 279, Thorvaldsens museum. Anden afdeling. Katalog, B 279, Thorvaldsens Museum, Haandkatalog for besøgende, J.L. Jensens penselskrift appliceret på positive kontrolmalerier Statens Museum for Kunst KMS 531, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 531, kartotekskort KMS 1063, Inventarium over Malerierne i Den Kgl. Malerisamling KMS 1063, kartotekskort KMS 1063, sagførers håndskrevne kopi af brev fra van Osten, uddrag J.L. Jensens penselskrift appliceret på andre malerier af J.L.Jensen Thorvaldsens Museum, Hvide liljer og grene af et rosentræ, B B 236, kartotekskort B 236, Charlottenborgs udstillingsliste B 236, Thorvaldsens Museum, Haandkatalog for besøgende, B236, Fortegnelse over malerierne og tegningerne i Thorvaldsens Museum A Museumsinspektør, kunsthistoriker, mag.art. William Gelius, mail B 236, Thorvaldsens Museum, katalog, B 236, Restaureringsrapport af konservator Steen Bjarnhof B 236, Tilstandsrapport af bevaringschef, konservator Henrik Bjerre Udtalelse fra Fælleskonserveringen, Center for Bevaring og Restaurering af Kunst Seniorforsker, mag.art. Marie-Louise Berner, mail Diskussion Statistiker, ph.d. Judith L. Jacobsen, mail

5 1. Indledning 1. Copydan, jurist Kirsten Kierkegaard, mail Fra: Kirsten Kierkegaard Sendt: 21. august :59 Til: 'Susan Ritterband' Emne: SV: dokumentation Kære Susan Ritterhand, Hermed min vurdering af retsstillingen i forbindelse med kopiering og signering af kunstværker: Kunstnerens eneret til eksemplarfremstilling (herunder håndværksmæssig kopiering)af hans værker Udgangspunktet er ophavsretslovens 2: Ophavsretten medfører, med de i denne lov angivne ændringer, eneret til at råde over værket ved at fremstille eksemplarer af det og ved at gøre det tilgængeligt for almenheden i oprindelig eller ændret skikkelse, i oversættelse, omarbejdelse i en anden kunstart eller i en anden teknik. Eneretten omfatter således ikke kun fremstilling af en identisk kopi af værket, men også fremstilling af en kopi af værket i ændret skikkelse f.eks. fremstilling af et postkort med fotografi af en skulptur eller fremstilling af en statuette af en skulptur i andet materiale og andet format end originalen. Eneretten omfatter også håndsværksmæssig kopiering af et eksisterende værk og dækker både det forhold, at der tilstræbes en stor grad af lighed med originalen og det forhold, at der tilføjes nogle mindre ændringer, jf. udtrykket i ændret skikkelse. Ophavsrettens beskyttelsestid I henhold til ophavsretslovens 63 stk. 1 varer ophavsretten til et værk indtil 70 år er forløbet efter ophavsmandens dødsår. Eksemplarfremstilling til privat brug Ophavsretslovens 12 indeholder følgende vigtige undtagelse fra den ophavsretlige eneret: Af et offentliggjort værk må enhver fremstille eller lade fremstille enkelte eksemplarer til sin private brug, såfremt det ikke sker i erhvervsøjemed. Sådanne eksemplarer må ikke udnyttes på anden måde. Særligt for så vidt angår kunstværker, hvorved forstås værker af billedkunst, bygningskunst og brugskunst samt fotografiske værker bestemmer 12 stk.2 nr. 2, at bestemmelsen i stk. 1 ikke giver ret til at fremstille et eksemplar af et kunstværk ved afstøbning, ved aftryk fra original plade eller stok eller på nogen anden måde, som indebærer, at eksemplaret kan opfattes som en original. Formuleringen, på nogen anden måde, som indebærer, at eksemplaret kan opfattes som en original omfatter alle former for eksemplarfremstilling, herunder reprografisk kopiering (f.eks. farvefotokopier i høj kvalitet) og håndværksmæssig fremstilling af eksemplarer. Håndværksmæssige kopieringer, hvor der ikke foreligger fuldstændig identitet med originalen er tilladt i studieøjemed, hvis det tydeligt angives, at der er tale om en eftergørelse. Hvis en kopi udformes på en sådan måde, at den fremtræder som en original, kan der efter omstændighederne straffes for mærkefalsk efter straffelovens 176, stk. 2. Eksemplarer af kunstværker, fremstillet til privat brug må ikke udnyttes på anden måde. Heri ligger, at en privatkopi ikke efterfølgende må spredes uden for den private familie- eller omgangskreds. En kopi, fremstillet til privat brug må f.eks. ikke sælges, bortgives eller udlejes. Den må heller ikke gøre til 5

6 genstand for offentlig fremførelse, f.eks. ved visning i en TV udsendelse eller en offentlig filmforevisning. Signering af kunstværker Ophavsretslovens 74 vedrører signering af kunstværker, dvs. angivelse af kunstnerens navn eller mærke (signatur mv.) på eksemplarer af kunstværker. Et pseudonym må antagelig sidestilles med et navn eller mærke. Bestemmelsen fastslår følgende: Stk. 1: På et kunstværk må kunstnerens navn eller mærke ikke anbringes af andre end kunstneren selv, medmindre denne har givet sit samtykke hertil. Stk.2: Kunstnerens navn eller mærke må ikke i noget tilfælde påføres et eftergjort eksemplar, således at det kan forveksles med originalen. Det er især stk. 2, der har interesse for den problemstilling, du rejser. Bestemmelsen beskytter kunstneren og almenheden imod, at kopier af originale kunstværker, herunder efterligninger, påføres kunstnerens navn på en sådan måde, at kopien kan forveksles med originalen. 74 gælder, uanset om værkets beskyttelsestid er udløbet. Foreligger der en ulovlig efterligning af et beskyttet værk (f.eks. en kopi, fremstillet med henblik på salg) hvor navnet på ophavsmanden til det efterlignede værk påføres værkseksemplarerne, er der sket en overtrædelse af ophavsretslovens 2 og evt. straffelovens 279 om bedrageri. (Husk, at der heller ikke til privat brug må fremstilles kopier, der kan opfattes som originaler). Er det efterlignede værk ikke beskyttet (beskyttelsestiden er udløbet), foreligger der ikke en overtrædelse af ophavsretslovens 2, men efter omstændighederne en overtrædelse af ophavsretslovens o Jeg håber, ovenstående har givet dig svar på dine spørgsmål. Ellers er du velkommen til at vende tilbage. Det er en vigtig problemstilling, du rejser, idet den har betydning for alle, der beskæftiger sig med håndværksmæssig kopiering af andres værker f.eks. til studiebrug. Med venlig hilsen Kirsten Kierkegaard 6

7 2. Kunsten som narrespeil, hjemmeside /july/kunsten_som_narrespeil Nettutgaven, onsdag 24. maj, 2006 Kunsten som narrespeil Galleri F15 i Moss viser i sommer malerier av Edvard Munch, Christian Krohg, Kai Fjell og Jakob Weidemann. Det skulle man i alle fall tro. Marte Stubberød Eielsen, Moss - MARTE.EIELSEN@KLASSEKAMPEN.NO, FREDAG 02. JULI, 2004.Strategi Men når sannheten bak forfalskningen har kommet for en dag, er det på tide å avsløre strategiene. Vi går inn i rommet med bilder til forveksling lik verk fra den norske gullalderen; Fritz Thaulow, Christian Krohg, Hans Heyerdahl og den svenske maleren Anders Zorn. Zorns bilder er egentlig malt av S. Jürgensen i S. Jürgensen malte aldri bildene som forfalskninger, men som kopi, forteller Hofsvang. Vi går nærmere et av bildene, hun peker på området rundt signaturen. - Ser du at malingen her har en litt annen valør og kvalitet enn resten av maleriet? Det er fordi kunstneren opprinnelig signerte med egen underskrift. Krohg-bildet er opprinnelig signert «Nach C. Krohg», altså «Etter C. Krohg». Dette er svært vanlig, at kunstnere i en læreprosess maler etter andre kunstnere. Hvor mange har ikke stått i Nasjonalgalleriet og malt etter «Brudeferden i Hardanger» for eksempel? Signaturen avgjør. - Forbrytelsen oppstår idet man kopierer kunstnerens signatur. Det er å regne som dokumentforfalskning, forteller konservatoren Johannes Rød, som har skrevet boka «Falsk», om kunstforfalskning. Forfalskning blir særlig attraktivt når en kunstners verk stiger i verdi, som Christian Krohgs bilder. - Krohg laget mange versjoner av sine egne bilder. Han hadde elevkretsen sin rundt seg, og nølte ikke med å signere elevenes verker, la elevene male litt på egne bilder, eller justere på elevenes verker, forteller Hofsvang. Denne praksisen har gjort det lettere å forfalske Krohg i ettertid. - Det er samme opphavsmann bak forfalskningen av både Krohg og Zorn fra vår utstilling, selv om det er ulike kunstnere som har malt bildene. Opphavsmannen var kunsthandler, han ble gjennomskuet på 1980-tallet og fikk fengselsstraff, forteller Hofsvang..Maleriet «Sommer» ble gjennomskuet på signaturen, som er langt enklere å avdekke enn de estetiske detaljene. - På avstand er «Sommer» svært lik en Munch.KunstVi kommer ikke utenom det evige spørsmål:- Men er det kunst?vi lar rådgiveren i politiet dømme i den saken:- En kopi uten forfalsket signatur: ja, det kunst. Men med forfalsket signatur: nei, det er falsk kunst, sier Jørgensen. Nettutgaven gir deg et lite utvalg fra dagens avis. For å lese hele avisen kan du abonnere. Klassekampen AS Ansv. red: Bjørgulv Braanen Daglig leder: Marga van der Wal Sentralbord: Besøk: Grønlandsleiret 31, Oslo Post: Boks 9257 Grønland, 0134 OSLO E-post: klassekampen@klassekampen.no Alt innhold er opphavsrettslig beskyttet Klassekampen Generelle vilkår for rettigheter til stoff som leveres til Klassekampen for publisering finner du her. 7

8 3. Todays Art Forgery, hjemmeside NOT painted by Henri Matisse Tuesday, 23 May, 2006, 22:33 Watch this item The seller, picasso-and-friends, Woodbury, New York, United States: I am offering this beautifully colored pastel and crayon on aged paper, signed H. Matisse in the lower right. Piece is titled "Yellow Flowers With Glass Decanter". The piece measures 11.3 inches by inches, unframed. Paper has uniformly tanned with age. Overall excellent condition- sold unframed. In addition to the Provenance (as listed below), the artwork comes with a Photo-Certificate of Authenticity on the signature by Herman Darvick. Piece may have been done while Matisse lived in Nice on the Cote dã 'Azur in the early 1920's. Provenance: (1)Private collection British government official., Lancashire, England. (2)Gifted to the son of the above, 1961, Lancashire, England (3)Sold by the above to present owner, Paperwork included with the piece: (1)Provenance as described above. Image of the piece within the Provenance document.. (2)Transfer of ownership document from present owner to the new buyer. (3)Signed Photo-Certificate of Authenticity on Matisse signature by Mr. Herman Darvick.(see additional info below) Addendum: (1)The British government official was an extensive collector of the arts until his death in This piece, along with other pieces in the collection, were acquired over a period of years during which he traveled extensively throughout Europe. It was known that several of the artists were personal friends. His son has certified that he received the collection as a gift when he started college in The collection has been in storage since that time, and as such the pieces are unframed and show normal aging and oxidation. (2)Mr. Herman Darvick is considered an expert in the field of signatures and autographs. He is author of "Collecting Autographs", Messner Books/Simon and Schuster,1981. He is author, "Autograph" entry, World Book Encyclopedia, He is former President of the Universal Autograph Collectors Club

9 Herman M. Darvick, Autographs, , Herman Darvick Autograph Auction, , Chief Cataloguer and Authenticator, Gallery of History, artfakes.dk: What a wonderful pack of lies. NOT a drawing by Picasso Monday, 22 May, 2006, 17:48 Watch this item The seller, gouyrand, West Coast, Canada: PICASSO DRAWING "LES BAIGNEUSES". WITH COA. Re listed due to non-payment by a "0" feedback winning bid A 33 x 25 cm drawing signed Picasso and dated September 15, The COA has wax seal with monogram applied, pristine condition, glazed. The same monogram wax seal is applied at the back of the frame. Low starting bid, considering the possible value of this piece. Overall size: 39.5 x 50 cm. Provenance U.K. artfakes.dk: The COA is a fake too. NOT painted by Picasso Sunday, 21 May, 2006, 14:53 Watch this item The seller, cahillart, California, United States: Offered for auction is this crayon/colored graphite drawing signed Picasso.the drawing is dated it is 9

10 in overall excellent condtion for its age. Guaranteed not a print of any kind. It has been stored unframed. The measurements are 12.6 inches wide and 9.2 inches high. It was acquired in france approximately thirty years ago at an estate art sale that contained a large collection of european and american fine art. The estate sale was held aboard a luxury yacht near cannes on the French riviera. An appraisal only, will acompany the drawing. Sold as is, (as described). Do your own research prior to bidding. No reserve price and auction results are final. artfakes.dk: Thousands of paintings, pastels and drawings signed Picasso have been offered for sale at ebay during the past ten years, but not a single of them was executed by him. NOT painted by the Danish interior painter Carl Holsöe Saturday, 20 May, 2006, 00:16 The seller, catherine_long, SHANGHAI, China: This is an oil paiting sunlit interior.the painting is in a very good condition.this is an oil painting on canvas signed"c.holsoe",a nice painting from the early 20th century.please visit all my other good oil paintings. artfakes.dk: This painting is for sure NOT Made in Denmark it's Made in China. A good copy, yes! But not good enough to be a genuine painting by Holsöe. 10

11 2. J.L. Jensen 4 Theobald Steins erindringer, Erik Steins afskrift, mail From: Erik Stein To: Susan Ritterband Sent: Tuesday, May 18, :39 PM Subject: I.L. Jensen Min far Torben Malling Stein skrev efter håndskrevet originalmateriale min oldefars erindringer på en gammel skrivemaskine omkring Vi har siden min fars død i 1972 haft en fotokopi i en ikke særlig læseværdig kvalitet liggende, men med computerens indtog var det nærliggende at skrive det hele nok en gang. En sådan afskrift kan på mange måder ikke blive bedre end kildematerialet, ikke mindst har datidens retskrivning voldt kvaler, men de værste afskriftfejl er blevet rettet således, at der ikke er ændret på det sprog, som blev anvendt dengang, min oldefar skrev disse erindringer i år I forhold til kildematerialet er der indføjet nogle få fodnoter, og mere end 200 slutnoter, omhandlende langt de fleste af de personer min oldefar har haft berøring med igennem sit liv som kunster, og som er nævnt i hans erindringer. Det være sig kongelige, adelige, kunstnere i mange afskygninger og andet godtfolk. For at sikre at disse noter er påhæftet de rigtige personer, er de krydschecket med oplysningerne i Gyldendals leksikon og Dansk biografisk leksikon (nævnt som Kilde 1 i slutnoterne) samt Weilbachs Kunstnerleksikon fra Sidstnævnte er i slutnoterne benævnt Kilde 2. Jeg er ikke bekendt med andre der har skrevet om Blomstermaleren Jensen, som han lidt respektløst betegnes i familien. Hvis du er interesseret må du gerne få en kopi af Theobald Steins Erindringer som min bror Torben har slutredigeret. De bedste hilsner Erik Stein Kære Susan Ritterband Som lovet sender jeg det uddrag fra min oldefar Theobald Steins erindringer der vedrører Blomstermaler Jensen ordret citeret: I 1855 var den første Verdensudstilling i Paris, og Fader vilde opmuntre mig ved en Rejse dertil i en Maanedstid. Jeg havde bestemt mig paa at rejse uden Selskab, men paa Skibet traf jeg Blomstermaler Jensen i[i] og portraitkunstner Gertner, og jeg blev ved Faders Præsentation knyttet til dem som Rejsefæller. I Paris kom vi til at bo i samme Hotel Garni, og jeg kom destoværre til at dele Værelse og Alkove med sidstnævnte. Naaer jeg udtaler mig saaledes, er det, fordi jeg led under hans udannede og cyniske Væsen, og han holdt mig længe vaagen, idet han holdt timelange Gardinprædikener om sine Meritter i Akademiet, hvoraf de fleste forekom mig meget malpropre. At komme ind herpaa vilde blive for vidtløftigt og ikke Umagen værd. L. Jensen var gammel og blev ubehageligt berørt ved at see sine Billeder paa Udstillingen mellem andre Blomsterbilleder, særlig Grønlands. ii[ii] 11

12 Jensen rejste hjem efter kort Tid, og snart efter bortkaldte Døden ham. Han havde været en af de virksomste Forsvarere i Akademiet for de conservative Anskuelser og med sine altid parate Talegaver, havde han kjæmpet imod Høyen og hans Parti, der mere og mere gjorde sig gældende imod Akademiet. Man mærkede ved flere Lejligheder - ogsaa i hans Kunst - Alderens Indflydelse. Den 31 Marts 1854 blev Akademiets 100 Aars Stiftelsesdag fejret ved en Fest paa Skydebanen. Selvom de reneste Høyensianere bleve hjemme, var dog Tilslutningen stor. Jensen skulde holde Talen for Akademiet. Han havde hjemme lært den Ord til Andet udenad og gik lige i Begyndelsen ganske fra Concepterne. Efter flere Gange at have taget fat igjen og stoppet strax igjen, sank han ned paa en Stol under Forsamlingens ubarmhjertige Latter og Fnisen, og Kammerassessor Collin iii[iii] sprang op paa sin Stol og meddelte Selskabet, at det var en Skaal for det Kgl. Akademi for de skjønne Kunstner, der nylig var bleven udbragt, og den blev saa drukket under megen Munterhed. 1[i] Johan Laurentz Jensen, , maler. Fik vejledning af Eckersberg, og lærte blomstermaleri af C. D. Fritzsch og valgte denne genre som speciale. Kilde 2. 1[ii] Theude Grønland, , maler. Udenlandsk medlem af Akademiet i Kbh Kilde 2. 1[iii] Formodentlig Jonas Collin, eller sønnen Edward Collin, Kilde 2. 12

13 5. Replikker og kopier af malerier af J.L. Jensen og af hans elever J.L. Jensen Privateje Skole 48 x 37 cm Bruun Rasmussen, 690 løbenr J.L. Jensen, usign. 26 x 38 cm. Privateje J.L. Jensen 12 x 16 cm. Bruun Rasmussen, 690 løbenr J.L. Jensen, x 36 cm. Bruun Rasmussen, 739 løbenr.1442 J.L. Jensen 26 x 36 cm. Bruun Rasmussen 734 løbenr

14 Sign. L.J. (J.L. Jensen) 17 x 27 cm. Bruun Rasmussen, 743, løbenr C. Moe, 1851, 23 x 31 cm. Bruun Rasmussen, 743, løbenr Skole 25 x 32 cm. Bruun Rasmussen, 750, løbenr J.L. Jensen, x 58,9 cm. Privateje Skole? cm. Antik Damgaard-Lauritsen. Mimi Seinsen? 20 x 15 cm. Bruun Rasmussen, 642 løbenr.343 Skole 58,4 x 42,8 cm. Thorvaldsens Museum B 236 J.L. Jensen 18 x 13 cm. Bruun Rasmussen, 729 løbenr

15 3. Litteratur om penselskriftsanalyse og tangerende analysemetoder 6. A digital technique for art authentication, artikel Published online before print November 24, 2004, /pnas PNAS December 7, 2004 vol. 101 no COMPUTER SCIENCES A digital technique for art authentication Siwei Lyu *, Daniel Rockmore *,, and Hany Farid *, Departments of * Computer Science and Mathematics, Dartmouth College, Hanover, NH Communicated by David L. Donoho, Stanford University, Stanford, CA, September 1, 2004 (received for review May 13, 2004) Abstract We describe a computational technique for authenticating works of art, specifically paintings and drawings, from high-resolution digital scans of the original works. This approach builds a statistical model of an artist from the scans of a set of authenticated works against which new works then are compared. The statistical model consists of first- and higher-order wavelet statistics. We show preliminary results from our analysis of 13 drawings that at various times have been attributed to Pieter Bruegel the Elder; these results confirm expert authentications. We also apply these techniques to the problem of determining the number of artists that may have contributed to a painting attributed to Pietro Perugino and again achieve an analysis agreeing with expert opinion. It probably was not long after people began paying money for art that a lucrative business in forging art was born, and it probably was not too much later that techniques for detecting art forgeries emerged. Even today, the early techniques for authentication remain preeminent. By and large, these techniques are based on "connoisseurship" and so rely on the discerning eyes of a few experts who are steeped in the work and life of the artist in question. Their opinion may be informed by the "catalogue raison," which is the current acknowledged authoritative work on the artist's oeuvre. Other desiderata may include provenance that might be traced back to the artist's circle or his collectors and makes possible the comparison of the work's implicit biography with the histories of related works or even a detailed analysis of any signature that may be present. (See ref. 1 for a survey of current techniques.) In addition to the reliance on the human actor, quantitative methods can be brought to bear. X- ray analysis can reveal a painting beneath a painting that can shed light on its origins. Surface analysis of the painting materials is another approach, most famously applied in the investigation of the famous "van Meerghen forgeries." In this case, the forgery of paintings 15

16 attributed to Johannes Vermeer was confirmed by dating the paintings according to the proportion of a certain lead isotope in the lead-based paint. An elementary application of differential equations allowed for the actual isotope content to be compared with the expected content had the work been painted in Vermeer's day (2). This technique marked a first use of mathematics in the service of art authentication. With the advent of powerful digital technology, computational tools may be able to provide new insights into and techniques for the art and science of art authentication. For example, a fractal analysis of Jackson Pollock's drip paintings has revealed interesting relations between the evolution of Pollock's aesthetic and the fractal dimension of his work (3, 4). The analysis also raises the possibility of using fractal dimension to help authenticate work attributed to Pollock. Various techniques from machine learning have been applied to the analysis and classification of "craquelure," the crack lines that appear over time in a painting (5). In this paper we present a computational tool for analyzing prints, drawings, and paintings for use in authentication. Specifically, we performed a multiscale, multiorientation image decomposition (e.g., wavelets) of a collection of high-resolution digital scans of a drawing or painting. This decomposition changes the basis from functions maximally localized in space (pixels) to one in which the basis functions are localized not only in space but also in orientation and scale. A familiar analogy comes from sound, where the original sound might be transformed into a vector of local frequency information that reflects how much of each frequency comprises the original sound over a short time window. We constructed a compact model of the statistics from such a multiscale, multiorientation image decomposition, and we looked for consistencies or inconsistencies across different drawings or paintings or within a single work. The latter is the so-called "problem of many hands" in which we try to determine the regions of a collaborative work that have been accomplished by a single artist. The analysis produced locally oriented spatial frequency data and so suggests that the accompanying model captures the subtle pen and brush strokes characteristic of an artist. Although an imitation may be perceptually similar to an original (i.e., very much in the "style of the master"), the subtle differences in stroke can reveal the presence of an imitation. In a sense, this work is a natural successor to the mathematical techniques used for graphology, or handwriting analysis (6), distilling not just the characteristic lines and curves of a painter's literal signature (which is often part of the process of authentication) but, even more, moving toward a characterization of the artist's aesthetic signature, resident within the line and curve of his or her work. Analogous techniques already have made their way into the literary world, where they fall within the discipline of "stylometry" (7). The problem of classification has been applied to divvying up the attribution of The Federalist Papers between James Madison and Alexander Hamilton (8) and the determination of the authorship of the 15th book in the Oz series (9). Statistical approaches to the question of authentication have surfaced in the analysis of William Shakespeare's sonnets (10). The problem of many hands finds its mirror in a study of the conjectured multiple authorship of the Old Testament (11). We began by applying our analysis to the problem of authentication of a collection of 13 drawings that have at one time been attributed to the famous draftsman Pieter Bruegel the Elder. We followed with a "many-hands" analysis of a portrait by the great Renaissance painter Perugino. We closed with a synopsis of the tools used in the analysis and described the underlying statistical model. We collected some of the more technical points in an appendix. 16

17 Materials Bruegel. The Flemish painter and draftsman Pieter Bruegel the Elder (1525/ ) was among the greatest artists of the 16th century. Of particular beauty and fame are Bruegel's landscape drawings. Over time he acquired many imitators; undoubtedly some simply were eager to work in the style of the great master, whereas others surely were hoping to pass off their work as Bruegel's for monetary gain. Some of these followers and imitators were expert enough that after being unmasked (or discovered) they became famous in their own right, e.g., Jacob Savery. Bruegel's work recently has been the subject of renewed study and interest (12 14). As a result, many drawings formerly attributed to Bruegel now are attributed to others (Table 1). View this table: [in this window] [in a new window] Table 1. Authentic and imitation works of art The delicate line and shading characterizing these works suggests that their classification according to a wavelet-like analysis may be both appropriate and fruitful. For our analysis we digitally scanned (at 2,400 dots per inch) eight authenticated drawings by Bruegel and five acknowledged Bruegel imitations from 35-mm color slides (Table 1), which were provided courtesy of the Metropolitan Museum of Art, New York (13). These color [red, green, blue (RGB)] images, originally 3,894 x 2,592 pixels, were cropped to a central 2,048 x 2,048-pixel region, converted to grayscale (gray = R G B), and autoscaled to fill the full intensity range [0,255]. Fig. 1 shows examples of an authentic drawing and an imitation. Fig. 1. Authentic works by Bruegel and imitations. An authentic work (Top) and an imitation (Middle) (nos. 6 and 7, respectively, in Table 1). (Bottom) The results of analyzing eight authentic Bruegel drawings ( ) and five imitations ( ). Note how the imitations lie significantly outside the bounding sphere of authentic drawings. View larger version (56K): [in this window] [in a new window] Each digital image then was subdivided into 64 nonoverlapping 256 x 256-pixel regions. Each of these subimages then was transformed by using a five-level, three-orientation wavelet-like decomposition (see Methods for details). From this decomposition, a 72 feature vector of coefficient and error statistics was extracted for each subimage (see Methods). Each drawing then corresponded to a set of points in this 72-dimensional space. Authentication was 17

18 indicated by the distance between these point clouds, with the belief that works by the same artist would be close together, irrespective of content, and that an imitation would be relatively far from the authenticated Bruegels. Thus, we first computed the Hausdorff distance (15) between all pairs of images (see Appendix A). The resulting 13 x 13 distance matrix then was subjected to a multidimensional scaling [(MDS) ref. 16] (see Appendix B). Fig. 1 shows the result of visualizing the original 13 images in a lower-dimensional space as determined by a MDS analysis. The circles in Fig. 1 correspond to the authentic drawings, and the squares correspond to the imitations. For purposes of visualization, the wire-frame sphere was rendered at the center of mass of the eight authenticated drawings and with a radius set to fully encompass all eight data points (in so doing, we assume knowledge of the authenticated Bruegels). Note that all five imitations fall well outside the sphere. The distances from the authenticated Bruegels to the center of the sphere are 0.34, 0.35, 0.55, 0.90, 0.56, 0.17, 0.54, and The distances from the imitations are considerably larger, at 1.58, 2.20, 1.90, 1.48, and 1.33 (the difference in the means of these two distance populations is statistically significant: P < 1 5, one-way ANOVA). Even in this space of reduced dimension, there is a clear difference between the authentic drawings and the imitations. Perugino. Pietro di Cristoforo Vannucci (Perugino) ( ) is well known as a portrait and a fresco painter, but perhaps he is best known for his altarpieces. By the 1490s, Perugino maintained workshops in Florence and Perugia, Italy, and he was quite prolific. Shown in Fig. 2 is the painting "Madonna with Child" by Perugino. As with many of the great Renaissance paintings, however, it is likely that Perugino only painted a portion of this work, and apprentices did the rest. To this end, we wondered whether we could uncover statistical differences among the faces of the individual characters. Fig. 2. How many hands contributed to this painting? (Upper) "Madonna with Child" by Perugino. (Lower) The results of analyzing the Perugino painting. The numbered data points correspond to the six faces (from left to right). Note how the three leftmost faces (1 3) cluster, whereas the remaining faces (4 6) are distinct. This clustering pattern suggests the presence of at least four distinct hands. View larger version (48K): [in this window] [in a new window] The painting (at the Hood Museum of Art, Dartmouth College) was photographed by using a large-format camera (8 x 10-inch negative) and drum-scanned to yield a color 16,852 x 18,204-pixel image. As in the previous section, this image was converted to grayscale. The facial region of each of the six characters was manually localized. Each face then was partitioned into nonoverlapping 256 x 256-pixel regions and autoscaled into the full-intensity range (0, 255). This partitioning yielded (from left to right) 189, 171, 189, 54, 81, and

19 regions. The same set of statistics described in the previous section was collected from each of these regions. Also as in the previous section, we computed the Hausdorff distance (see Appendix A) between all pairs of six faces. The resulting 6 x 6 distance matrix then was subjected to MDS (see Appendix B). Fig. 2 shows the result of visualizing the original six faces in a lower-dimensional space as determined by a MDS analysis. The numbered data points correspond to the six faces (from left to right) in Fig. 2. Note how the three leftmost faces cluster, whereas the remaining faces are distinct. The average distance between faces 1 3 is 0.61, whereas the average distance between the other faces is This clustering pattern suggests the presence of at least four distinct hands and is consistent with the views of some art historians (T. Barton Thurber, Hood Museum, Dartmouth College, personal communication). Methods Our methodology makes use of a decomposition of images by using basis functions that are localized in spatial position, orientation, and scale (e.g., wavelets). These sorts of expansions have proven extremely useful in a range of applications (e.g., image compression, image coding, noise removal, and texture synthesis). One reason for this utility is that such decompositions exhibit statistical regularities that can be exploited (17 19). Described below is one such decomposition and a set of statistics collected from this decomposition. The decomposition is based on separable quadrature mirror filters (20 22). As illustrated in Fig. 3, this decomposition splits the frequency space into multiple scales and orientations. This decomposition is accomplished by applying separable low-pass and high-pass filters along the image axes, generating a vertical, horizontal, diagonal, and low-pass subband. For example, the horizontal subband is generated by convolving with the high-pass filter in the horizontal direction and the low-pass filter in the vertical direction, the diagonal band is generated by convolving with the high-pass filter in both directions, etc. Subsequent scales are created by subsampling the low-pass subband by a factor of two and recursively filtering. The vertical, horizontal, and diagonal subbands at scale i = 1,..., n are denoted as V i (x, y), H i (x, y), and D i (x, y), respectively. Fig. 4 Right and Left shows a three-level decomposition of the scanned Perugino work, respectively. Fig. 3. An idealized multiscale and orientation decomposition of frequency space. From top to bottom, levels are 0, 1, and 2. From left to right are low-pass, vertical, horizontal, and diagonal subbands. View larger version (12K): [in this window] [in a new window] 19

20 Fig. 4. The absolute values of the subband coefficients at three scales and three orientations (the residual lowpass subband is shown in the upper left corner) (Right) for the Perugino painting (Left). View larger version (82K): [in this window] [in a new window] Given this image decomposition, the statistical model is composed of the mean, variance, skewness, and kurtosis of the subband coefficients at each orientation and at scales i = 1,..., n 2. These statistics characterize the basic coefficient distributions. To capture the higherorder correlations that exist within this image decomposition, these coefficient statistics are augmented with a set of statistics based on the errors in an optimal linear predictor of coefficient magnitude. As described in ref. 19, the subband coefficients are correlated to their spatial, orientation, and scale neighbors. For purposes of illustration, consider first a vertical band, V i (x, y), at scale i. A linear predictor for the magnitude of these coefficients in a subset of all possible neighbors may be given by where w k denotes scalar weighting values, and denotes magnitude. This particular choice of spatial, orientation, and scale neighbors was used in our earlier work on detecting traces of digital tampering in images. ** Here, we employed an iterative brute-force search (on a persubband and per-image basis) for the set of neighbors that minimizes the prediction error within each subband. Consider again the vertical band, V i (x, y), at scale i. We constrained the search of neighbors to a 3 x 3 spatial region at each orientation subband and at three scales, namely, the neighbors [1] 20

21 with c x = { 1,0,1} and c y = { 1,0,1}, and, of course, excluding V i (x, y). From these 80 possible neighbors, the iterative search begins by finding the single most predictive neighbor [e.g., V i+1 (x/2 1, y/2)]. This neighbor is held fixed, and the next most predictive neighbor is found. This process is repeated five more times to find the optimally predictive neighborhood. On the kth iteration, the predictor coefficients (w 1,..., w k ) are determined as follows. Let the vector contain the coefficient magnitudes of V i (x, y) strung out into a column vector, and the columns of the matrix Q contain the chosen neighboring coefficient magnitudes also strung out into column vectors. The linear predictor then takes the form [2] where the column vector minimizing the quadratic error function:. The predictor coefficients are determined by [3] This error function is minimized by differentiating with respect to [4] setting the result equal to zero and solving for to yield [5] The log error in the linear predictor then is given by Once the full set of neighbors is determined, additional statistics are collected from the errors of the final predictor, namely, the mean, variance, skewness, and kurtosis. This entire process is repeated for each oriented subband, and at each scale i = 1,..., n 2, where at each subband a new set of neighbors is chosen and a new linear predictor is estimated. For an n-level pyramid decomposition, the coefficient statistics consist of 12(n 2) values, and the error statistics consist of another 12(n 2) values, for a total of 24(n 2) statistics. These values represent the measured statistics of an artist's style and are used to classify or cluster drawings or paintings. [6] 21

22 As stated above, after the computation of the feature vectors, MDS was used to project the original 72-dimensional feature vectors into a three-dimensional subspace. Features with no discriminating power (e.g., the means) therefore will play no role in the lower-dimensional embedding. Discussion We have presented a computational tool for digitally authenticating or classifying works of art. This technique looks for consistencies or inconsistencies in the first- and higher-order wavelet statistics collected from drawings or paintings (or portions thereof). We showed preliminary results from our analysis of 13 drawings either by or in the style of Pieter Bruegel the Elder as well as a painting by Perugino. We expect these techniques, in collaboration with existing physical authentication, to play an important role in the field of art forensics. Appendix A: Hausdorff Distance The Hausdorff distance is a distance metric defined on two sets of vectors, X and Y. The metric, H(, ) is defined as where h(, ) is defined as Here d(, ) can be any distance metric defined on the vector space subsuming X and Y. In our case, we use Euclidean distance d. Appendix B: MDS MDS is a popular method to visualize high-dimensional data. Given n vectors, where, the goal of MDS is to find a lower-dimensional embedding for these data that minimally distorts their pairwise distances. Denote the n x n distance matrix as, where d(, ) is a distance metric in. The most common such metric is Euclidean distance defined as. Given the pairwise symmetric distance matrix, the classic (metric) MDS algorithm is given by the following steps: 1. Let. 2. Let B = HAH, where, I n is a n x n identity matrix, and each component of the n-dimensional vector is Compute the eigenvectors,, and corresponding eigenvalues, 1,..., n, of matrix B, where n. 4. The new, lower-dimensional representation of the original data,, are then given by where denotes the ith component of the vector, and in our examples m' = 3. Acknowledgements 22

23 We thank George Goldner and Nadine Orenstein of the Metropolitan Museum of Art and Kathy Hart of the Hood Museum of Art for their assistance and cooperation. D.R. was supported by Air Force Office of Scientific Research Grant F H.F. was supported by an Alfred P. Sloan Fellowship, National Science Foundation CAREER Grant IIS , Department of Justice Grant 2000-DT-CS-K001, and departmental National Science Foundation Infrastructure Grant EIA Footnotes Author contributions: S.L., D.R., and H.F. designed research; S.L. and H.F. performed research; S.L., D.R., and H.F. analyzed data; and D.R. and H.F. wrote the paper. Abbreviation: MDS, multidimensional scaling. Henceforth we will give the benefit of the doubt to the imitator and use the term "imitation" rather than the more charged "forgery." Although converting from color to grayscale results in a significant loss of information, we did so to make it more likely that the measured statistical features and subsequent classification were more likely to be based on the artist's strokes and not on simple color differences. We also have experimented with both Laplacian and steerable pyramid decompositions. Results from a steerable pyramid (with eight orientation subbands) were similar to the results included above (which use only three orientation subbands). Furthermore, the Laplacian pyramid generally gave poor results. So, although it seems that oriented subbands are necessary, it also seems that a finer tuning of orientation is not necessary for this particular task. ** Farid, H. & Lyu, S. IEEE Workshop on Statistical Analysis in Computer Vision (in conjunction with Computer Vision and Pattern Recognition), June 17 23, 2003, Madison, WI. Integer rounding is used when computing the spatial positions of a parent, e.g., x/2 or x/4. To whom correspondence should be addressed at: 6211 Sudikoff Laboratory, Department of Computer Science, Dartmouth College, Hanover, NH farid@cs.dartmouth.edu by The National Academy of Sciences of the USA References 1. Spencer, R. D. (2004) The Expert Versus the Object (Oxford Univ. Press). 2. Keisch, B. (1968) Science 160, [ISI] 23

24 3. Taylor, R., Micolich, A. P. & Jones, D. (1999) Nature 399, 422.[Medline] 4. Taylor, R. (2004) Chaos Complexity Lett., in press. 5. Abas, F. S. & Martinez, K. (2003) Proc. SPIE Int. Soc. Opt. Eng. 5011, McKeague, I. W. (2004) J. Am. Stat. Assoc., in press. 7. Holmes, D. I. & Kardos, J. (2003) Chance 16, Mosteller, F. & Wallace, D. L. (1984) Applied Bayesian and Classical Inference: The Case of the Federalist Papers (Springer, New York). 9. Binong, J. N. G. (2003) Chance 16, Thisted, R. & Efron, B. (1987) Biometrika 74, [ISI] 11. Friedman, R. (1997) Who Wrote the Bible? (Harper, San Francisco). 12. Mielke, U., ed. (1996) Pieter Bruegel, Die Zeichnungen (Brepols, Turnhout, Belgium). 13. Orenstein, N. M. (2001) Pieter Bruegel the Elder (Yale Univ. Press, New Haven, CT). 14. Orenstein, N. M. (2003) Int. Found. Art Res. J. 6, Huttenlocher, D. P., Klanderman, G. A. & Rucklidege, W. J. (1993) IEEE Trans. Pattern Anal. Machine Intell. 15, [CrossRef][ISI] 16. Cox, T. & Cox, M. (1994) Multidimensional Scaling (Chapman & Hall, London). 17. Shapiro, J. (1993) IEEE Trans. Signal Processing 41, [CrossRef][ISI] 18. Rinaldo, R. & Calvagno, G. (1995) IEEE Trans. Image Processing 4, [CrossRef][ISI] 19. Buccigrossi, R. W. & Simoncelli, E. P. (1999) IEEE Trans. Image Processing 8 (12), [CrossRef][ISI] 20. Vaidyanathan, P. P. (1987) IEEE ASSP Magazine 4, Vetterli, M. (1987) IEEE Trans. ASSP 35, Simoncelli, E. P. & Adelson, E. H. (1990) Subband Image Coding (Kluwer Norwell, Boston). 24

25 7. The Authentic Project, hjemmeside Delft University of Technology (TU-Delft) Faculty of Electrical Engineering, Mathematics, and Computer Science (EEMCS) Information and Communication Theory Group The main research question of the project reads: "How can existing and new techniques from digital image processing and knowledge discovery (data mining) be developed and applied in order to determine the authenticity and dating of visual art?" The AUTHENTIC project replaces the subjectivity of the human expert by the objectivity of intelligent image-analysis algorithms that mimic the human visual-perception and interpretation process. It aims at incorporating knowledge about the way visual art data is processed by human expert in their art analysis into the computerised art analysis. The AUTHENTIC project is a computational realization of "Pictology", a checklist of aspects relevant in their assessment of the authenticity and dating of works of art proposed by Van Dantzig (Pictology: an analytical method for attribution and evaluation of pictures, Leiden, Brill, 1973.) However, it is not a literal translation of the original approach, but emphasizes more the underlying methodology. Analysis is performed at various levels. Varying from local features like lines, colour, texture, brush stroke et cetera, to high-level features like style, composition et cetera. Data mining techniques are used in order to become more insight into the relevance of features and other characteristics METHOD OF RESEARCH The major steps in the project are the following: making the pictorial features measurable [image analysis toolbox]; forming a large database with paintings; apply data mining techniques to induce concepts, models, associations; evaluate intermediate results and adjust measurements, models and accordingly the prediction rules; complete the project with an authenticity and dating toolbox and provide a performance analysis; Clearly, even if the net result is an useful toolbox for the art expert, there are two fundamental issues and research subjects at hand: PICTOLOGY: transforming the visual pictorial characteristic elements into measurable identities in the digitised image domain in such a way that the human art expert is still capable to perform his/her authenticity analysis; DATA MINING: induce low-level, intermediate-level and high-level concepts, models and associations that are understandable and explainable by human experts. Created at Wednesday, 30 June 2004 Last Updated Saturday, 09 October

26 8. Bevaringschef, konservator Jørgen Wadum, mail From: Jørgen Wadum Sent: 19. januar :55 To: Susan Ritterband Subject: SV: Vedr. mit speciale om J.L. Jensen Kære Susan, Tak for din mail. Så vidt jeg ved, er der ikke skrevet noget decideret om Rembrandts penselskrift. Det er vel snarere et spørgsmål om connoisseurship - ligesom det du har udviklet for Jensens malerier. Morelli metodik lever derfor på stadig i bedste velgående. Jeg vil derimod altid hævde, at ingen kunstkender kender dagen før naturvidenskaberne også har ladet deres lys skinne over genstanden. eg er meget begejstret for brugen af IR-imaging, ikke kun for at lede efter undertegninger, men snarere for at dokumentere undermaling og skitseagtige kompositionsstrøg. I denne proces kunne man nemlig formode, at kunstneren arbejde ud fra en vis automatik, hvorimod det færdige farvelag er meget mere maniera. Jeg er bange for, at mit svar ikke gav dig megen hjælp.... bedste hilsener, Jørgen 26

27 Penselskriftsanalyse af J.L. Jensens malerier. Autenticitet og forfalskninger 9. Van Dantzigs liste over Van Goghs karakteristika M.M. van Dantzig, Vincent? A new method of identifying the artist and his work and of unmasking the forger and his products, Keesing, Amsterdam, u.år. Løst indstik i bogen. 27