Framework for Building Interactive Plotting and Visualization Widgets

Performing quantitative image-based studies of biological specimens across multiple scales requires data-driven approaches to analyze the complex multi-dimensional feature space extracted from pixels, segmented subcellular objects, single cells, and multicellular structures. Interactive visualization and plotting of multivariate features and their relationships represents a crucial first step in this type of work. It provides insights into the quality of newly developed and applied image processing pipelines, reveals various types of structure in the obtained datasets, and is essential in guiding biologists in hypothesis generation and making unexpected observations. However, the field lacks tools that can interactively link feature visualization with the underlying image data.

For this reason, the team will build a flexible, easy-to-use and library-agnostic framework for building interactive plotting tools for the napari community. This framework will enable plugin developers to build widgets that can link images to tools that analyze the multivariate feature space extracted from those images. The group will establish the necessary framework to make this feasible and provide well-documented templates for making interactive plots. The project will create a general framework for plot visualization and consists of three parts: design, implementation, and deployment.

This framework will support many plugin developers by creating the basis for an interactive plotting set up and developing templates for interactive data visualization. With the overall strengthened plugin system and the recent push into feature extraction within napari, interactive feature visualization will progressively gain more importance. The seamless integration of future plugins will allow users to validate results in their images and build trust in napari-based image quantification workflows.