Deciphering the Spatial Location of Paintings in the ‘Kunsthaus Hamburg’: An Algorithmic and Philosophical Exploration Using Blurred Images

This study reconstructs the spatial layout of the Kunsthaus Hamburg using blurred photographs and transparent originals to gain insights into memory, perception, and cognition. We investigate mental models and visual representations during museum navigation in the 2009 and 2019 visits by employing a philosophical behavioural approach based on vignettes’ logics—scenarios exploring cognitive processes in constructing hypothetical stories. The study examines factors influencing deviations from planned routes, cognitive mechanisms behind selecting perspectives in photographs, and environmental awareness at image capture.

A hierarchy way of finding the orientation (HWO) algorithm breaks down the space into a hierarchy of nodes (rooms) and edges (pathways), determining the initial position within the museum based on the first clear photograph and the known layout. A hierarchical map categorizes nodes and edges by importance and connectivity, with major sections forming the top level and individual rooms forming the lower levels. Random forests classify images of rooms and predict the sequence of rooms visited based on features extracted from clear photographs. A labelled dataset of fifty room photographs was used to train the random forest model, which then classified the blurred photographs and combined this with the HWO algorithm to predict the sequence of rooms visited.

The study acknowledges biases such as inaccurate architectural measurements and the Halo effect. Insights from Rehkämper (2006) and Guilherme (2019) were applied to understand the interplay between mental models, expectations, and new information. This framework for reconstructing spatial layouts from blurred photographs provides insights into the spatial cognition and the dynamic nature of mental models.