Exhibition design conjoins distinct architectural and curatorial requirements. It is proposed that the common language of architecture and curatorship is space: how displays are arranged to be viewed in particular sequences and visual frames, placed in fields of co-visibility or grouped according to their spatial arrangement as well as their stylistic, historical or other classificatory labels. As visitors become immersed in exhibition space they are exposed to an informally staged pedagogy aimed at enhancing their enjoyment and understanding of the exhibition. The second floor of the High Museum of Art, with the permanent collection of objects, opened in 1983, is chosen as a case study. Meier designed the original building and decorative arts exhibition. Scogin and Elam produced a significant modification in 1997 to house a thematic exhibition. Lord Aeck and Sargent restored a simplified version of the original layout in 2003. Rigorous quantitative analyses document these successive changes and identify the fundamental shifts in exhibition design principles that they represent. Visual relationships, the break up of space and patterns of movement are analyzed using standard space-syntax methodologies. New techniques are proposed in order to describe and quantify overlapping patterns of spatial grouping. It is shown that the original design encouraged visitors to view and compare objects in alternative ways, generating open-ended readings and multiple understanding. The 1997 layout dictated sequences of viewing and framed frontal views in order to communicate how art engages human experience, including the body or the environment. The 2003 layout re-instated multiple viewing points and comparative groupings while emphasizing the individual work. The dissertation examines how architecture and curatorship interacted in a unique building which provides great experiential richness as well as design constraints. In addition, it demonstrates how descriptive theory can help bridge between architectural and curatorial intents by capturing the principles of arrangement which are fundamental to both.

As the demand for 3D maps from LIDAR scanners increases, delivering high-quality maps becomes critical. One way to ensure the quality of such maps is through point cloud alignment classification, which aims to classify the alignment error between two registered point clouds. Specifically, we present the classifier FACT (Feature-Aware Classification Transformer), consisting of two main modules: feature extraction and classification. Descriptive features are extracted from the joint point cloud, which are then processed by a point transformer-based neural network to predict the alignment error class. In a ten-class point cloud alignment classification test, FACT achieved 92.4% accuracy, where the alignment error ranged from zero meters and radians to 0.9 meters and 0.09 radians. Remarkably, the classifier only made one misprediction beyond neighboring classes, exhibiting its ability to detect alignment errors as the classes have an inherent order. Furthermore, when benchmarked on two binary classification tasks, FACT showed significantly superior performance over the baseline and even obtained 100.0% accuracy for the easier of the two tasks. FACT not only detects potential errors in 3D maps but also estimates their magnitude, leading to more reliable 3D maps with quality estimations for each transformation.