With the advent of new data
acquisition technologies, it is becoming increasingly common for art
museums to compile digital archives of their collections for
internal use. These records can include extremely high resolution
images of paintings or drawings under visible light, X-rays,
multispectral images (from which local paint pigment composition
information can be derived), or even laser scans of a paintings
three-dimensional surface. However, thus far, little attempt has
been made to subject this wealth of data to sophisticated
computational analyses, even though such analyses could potentially
address a variety of questions of great art historical interest.
Having backgrounds from different
disciplines like Applied Mathematics, Computer Science and
Electrical Engineering, we are using state of the art machine
learning and signal processing techniques to extract useful
information from digital scans of art works.