Cave art digital analysis subject of IST research

IST Professor James Wang returned recently from Firenze, Italy recently where he presented a paper titled “Determining the Sexual Identities of Prehistoric Cave Artists using Digitized Handprints -- A Machine Learning Approach” at the annual ACM International Conference on Multimedia.

Wang wrote the paper in collaboration with Weina Ge, Dean Snow, Prasenjit Mitra and C. Lee Giles.

The group’s paper was about using image processing and machine learning techniques to study prehistoric cave artists. Specifically, they teach computers how to recognize male and female handprints so it can determine the origin of specific examples, Wang said.

“Some of the cave arts were over 10,000 years old,” he said. “It is amazing that we can use modern tools to help reveal secrets from ancient time.”

The event took place from Oct. 25 to 29 at multiple conference centers in Firenze. One of the most prestigious conferences in the industry, more than 800 people in academia and research-oriented fields attended. Wang said he’s been participating in the conference since the 1990s and has had multiple works presented there.

“Most of our best work got published first in this conference before appearing in journals,” he said.

The group’s paper was featured in the cultural heritage track of the conference, a new area and one Wang said is starting to gain more attention from researchers in multimedia. It also “represents an ongoing interdisciplinary collaboration between IST and anthropology,” he said.

Anthropology Professor Dean Snow said his work on the project was funneled through research in Cyprus and the Cyprus Institute on Byzantine artwork analysis.

“What they’re interested in and trying to figure out is if there is some way to do an automated computerized analysis of painting that would produce information art historians haven’t been able to figure out in the traditional way they do research,” Snow said.

To read the paper, visit http://infolab.stanford.edu/~wangz/project/imsearch/hand/ACMMM2010/ online.

Last Updated July 07, 2011