Object Detection

Detection of artworks in installation views of the Museum of Modern Art, New York.

Frontier (2013-2015): “COMPOSITO Arthistoric Analysis of Architecture via Computer Vision”

MWK (2012-2015): “Objects in Mirror are Closer than they Appear: Objectrecognition and Automatic Comparison of the Sachsenspiegel Codices”

FRONTIER (2010-2012): “Computer-Assisted Detection and Analysis of Medieval Legal Gestures”

“Visual Recognition in Datasets of Pre-Modern Images”

Finding objects, such as cars, bottles, glasses, hats etc. in large image collections, has been a common task in computer vision. The efficiency and accuracy of algorithms to retrieve objects or selected image regions from digital images, has lead to their application to art data. Data sets have been evaluated, studying recurring motifs over time and space, thus helping to reveal artistic relations and general commonalities in art.

The group has performed object detection on various data sets, including paintings of crucifixion, architectural prints or medieval manuscripts to assist with art historical research. The following site provides an overview over done work and demonstrates the efficiency of computational tools to detect regions and eventually to evaluate art.

Click on the left images to get more information on our projects.