'Serendipity Searcher'

A project within BIO26 - Biennial of Design 2019 in Ljubljana/Slovenia.

Come and visit the 'Seredipity Searcher' until
9. February 2020 at the National Library NUK in Ljubljana! For more information about BIO26 check at


Serendipity Searcher is a visual atlas and a physical search engine that generates unexpected pathways to the valuable archive of the National and University Library (NUK). It brings together a spatial installation, machine learning, and human agency to physically reveal, select, and search for visual relations among the library’s holdings. The result is an ever-changing atlas of items that might otherwise never meet. The installation consists of a large light panel and transparent filters that can magically reveal images when placed within its frame. Visitors are invited to interact and physically move the filters to explore visual connections or generate new ones by highlighting any two images of interest. Inspired by the architectural ideals of Jože Plečnik, the project seeks to present an array of stories to a public curious enough to search for them.

So what is the Serendipity Searcher?

Technically speaking: It's a device made from 4 big LCD screens. The screens appear blank white unless you are viewing through movable transparent panes. These (polarization) filters allow you to see a selection of graphical content from the library. Starting with a random set of 16 pictures you can select 2 images by placing the tinted 'selection filters' on the ledges in front of the images. Now a machine learning algorithm will present a new set of images that connects the selected images in a visual way based on graphical features of the digital images. This is a kind of aesthetic pathway that allows you to find new connenctions. By removing the selection filters from the ledges a new set of 16 random images will appear and you can 'browse' and explore fresh images! 

How does it work?

The physical part is quite simple. We are using the very nature of a LCD (Liquid Crystal Display) which works with thin polarization filters that interfere with the light waves passing through the LCD module. Without the filters no color/contrast can be seen (so we removed the filters from the TV screens and put it on the panes instead) 
The graphical content is processed by 3 Raspberry Pi microcomputers. They run an algorithm, trained with about 10.000 images from the library. It analysizes the features of the images (e.g. colors, contrast, size, patterns,..) and creates connections between images.
The selection/trigger works with hall sensors inside the ledges that respond to magnets in the panes of the selection filters.


Thomas Hügin (Germany)

is an engineer and designer. He has a Diploma degree in product engineering from Furtwangen University of Applied Sciences and an MA in product design from the University of Edinburgh. He works as a mechanical design engineer and is the founder of Design Studio Platypus.

Maja Kolar (Croatia)

is an independent designer, researcher, and educator. She has an MFA degree from Konstfack University of Arts, Crafts and Design in Stockholm. Kolar acts as a creative director, head designer, and researcher in the Oaza design collective in Zagreb, which she cofounded.

Yuxi Liu (China)

is a multidisciplinary designer and researcher with backgrounds in Industrial Design, Design Informatics, and Interaction Design. Her work examines social, cultural, and ethical implications of emerging technologies through critical, philosophical, and ecological lenses. Her recent research focuses on designing with machine intelligence from a more-than-human-centered perspective.

Boris Smeenk

studied Digital Craft in Willem de Kooning Academy. With Arthur Boer, he started a practice based studio in Rotterdam, focusing on the contemporary digital culture (artificial intelligence, computer vision and human-machine dynamics). At the centre of their research, led by technological experiments, is the relation between algorithmic statistics and the human imagination.