The project shows a large image collection in an interactive 3D-visualisation. About 2300 prints and drawings from “Schweizer Kleinmeister” from the Gugelmann Collection of the Swiss National Library form a cloud in the virtual space.
The images are grouped according to specific parameters that are automatically calculated by image analysis and based on metadata. The goal is to provide a fast and intuitive access to the entire collection, all at once. And this not accomplished by means of a simple list or slideshow, where items can only linearly be sorted along one axis like time or alphabet. Instead, many more dimensions are taken into account. These dimensions (22 for techniques, 300 for image features or even 2300 for descriptive text analysis) are then projected onto 3D space, while preserving topological neighborhoods in the original space.
The project renounces to come up with a rigid ontology and forcing the items to fit in premade categories. It rather sees clusters emerge from attributes contained in the images and texts themselves. Groupings can be derived but are not dictated.
The user can navigate through the cloud of images. Clicking on one of the images brings up more information about the selected artwork. For the mockup, three different non-linear groupings were prepared. The goal however is to make the clustering and selection dependent on questions posed by any individual user. A kind of personal exhibition is then curated, different for every spectator.
Sonja Gasser, Nikola Marincic, Jorge Orozco and Mathias Bernhard
Crawling, extraction, image processing, machine learning:
Python (BeautifulSoup, sklearn, skimage)
Places search for Lat/Lng coordinates:
- <weblese> on SRF (Swiss National Radio and TV)
- Interview in “Kultur kompakt” on SRF2Kultur (starting 5:50)
- Article in Wikipedia:Kurier (scroll down a bit)
- Article on opendata.ch
- MIGROS Kulturprozent (event support)