WILL COM­PUT­ERS BE­COME PER­FECT ART FORG­ERS?

World of Knowledge (Australia) - - Technology -

The choice of artist was no co­in­ci­dence. “Rem­brandt com­pleted a lot of paint­ings, so there’s a huge amount of tech­ni­cal data avail­able,” ex­plains Joris Dik of Delft Univer­sity of Tech­nol­ogy. Com­put­ers can only learn the style of par­tic­u­larly pro­duc­tive artists – and even then they’ll not be 100% per­fect. Cer­tain de­tails, like the re­flected light from the col­lar onto the chin, are miss­ing from the im­age. “Still,” ex­plains art his­to­rian Gary Schwartz, “the tech­nique could help un­mask forg­eries from newly dis­cov­ered paint­ings.” Fu­ture soft­ware would com­pare all ‘real’ noses painted by Rem­brandt with the struc­ture of the new one and eval­u­ate their sim­i­lar­ity.

WORKS OF ART MADE FROM DATA This ‘man in the hat’ brings to­gether more than 168,000 im­age de­tails from hun­dreds of Rem­brandt paint­ings. It’s not just a sim­ple flat im­age ei­ther; depth and tex­ture was added to cre­ate a bona fide three-di­men­sional model (see be­low). The im­age was painted us­ing a 3D printer that used 13 lay­ers of ink. 0.05mm 0.1mm 0.18mm 0.2mm MOUN­TAINS OF COLOUR Re­gard­less of whether it’s printed or painted, ev­ery brush­stroke cre­ates peaks and val­leys on the can­vas. Some are so tiny they mea­sure less than one mil­lime­tre.

Newspapers in English

Newspapers from Australia

© PressReader. All rights reserved.