DROPLETS: A Marker Design for visually enhancingLocal Cluster Association
Year: 2024
Type: article
Abstract: The objective of the Redesign Challenge of the Bio+MedVis Challenge @ IEEE VIS 2024 is to redesign an existing visualization of multi-cell gene expressions of tissue samples. In this, multiple cells are accumulated into pixels. For each pixel the visualization should convey the prevalence and extent of cell types it is composed of, i.e., a proportional relation. The provided baseline technique of superimposed Pie charts -- a common technique for this kind of relation -- is not an ideal choice as the cell-type quantities of neighboring pixels are hard to compare due to a spatial disarray inherent to pie charts. This limits the perception of regions with coherent cell-type compositions, which constitutes one of the essential visual analytics tasks. We propose a novel marker design: \emph{Droplets} -- a space-saving design for visually enhancing the presence of clusters and regional borders. We evaluate this concept for the given tissue sample and compare it to the given baseline and other alternatives.
Institution: Graz University of Technology
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
Funder: European Commission
OpenAlex ID: W6911755512
FWCI:
Cites:
Cited by:
Related to: