Dutch researcher Laurens van der Maaten has developed an analytical technique that enables computers to better interpret and recognize the content of photos, images, and data. Van der Maaten has used his technique to distinguish between real and forged copies of paintings by Van Gogh, and from paintings by Van Gogh's contemporaries. The system also has been used to identify pottery, Roman coins, and glass from the Middle Ages.
A major challenge Van der Maaten had to deal with was the large number of pixels in these images, and the resulting high dimensionality of image-space representations. The variation of images that results from changes in illumination, rotations, or changes in scale also was a significant hurdle. These problems were solved by testing new techniques in visualization experiments and extrapolating those techniques and re-testing them with numerous variations.
The technique was developed for the automatic analysis in the field of cultural heritage, and can be used for the analysis of coins, seeds from archaeological excavations, or historic paintings. The system also can be applied to non-visual collections of high-dimensional data.
From Netherlands Organization for Scientific Research
View Full Article
Abstracts Copyright © 2009 Information Inc., Bethesda, Maryland, USA