Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

TIAViz: A Browser-based Visualization Tool for Computational Pathology Models

Mark Eastwood, John Pocock, Mostafa Jahanifar, Adam Shephard, Skiros Habib, Ethar Alzaid, Abdullah Alsalemi, Jan Lukas Robertus, Nasir Rajpoot, Shan Raza, Fayyaz Minhas

TL;DR

TIAViz addresses the need for flexible, browser-based visualization of digital pathology model outputs to support research and collaborative evaluation. Built as part of the TIAToolbox, it uses a Flask tile server and an SQLite-based AnnotationStore to render zoomable WSIs with layered overlays (annotations, segmentations, heatmaps, graphs) via Bokeh. The tool supports multiple overlays, graph visualizations, interactive AI workflows (e.g., HoVer-Net and potential GPT-vision integration), and enables remote or server-based demos. This open-source solution simplifies integration into Python-based workflows and enables rapid exploration of model outputs, with demonstrated applications in IGUANA, SNA, HiGGsXplore, and MesoGraph.

Abstract

Digital pathology has gained significant traction in modern healthcare systems. This shift from optical microscopes to digital imagery brings with it the potential for improved diagnosis, efficiency, and the integration of AI tools into the pathologists workflow. A critical aspect of this is visualization. Throughout the development of a machine learning (ML) model in digital pathology, it is crucial to have flexible, openly available tools to visualize models, from their outputs and predictions to the underlying annotations and images used to train or test a model. We introduce TIAViz, a Python-based visualization tool built into TIAToolbox which allows flexible, interactive, fully zoomable overlay of a wide variety of information onto whole slide images, including graphs, heatmaps, segmentations, annotations and other WSIs. The UI is browser-based, allowing use either locally, on a remote machine, or on a server to provide publicly available demos. This tool is open source and is made available at: https://github.com/TissueImageAnalytics/tiatoolbox and via pip installation (pip install tiatoolbox) and conda as part of TIAToolbox.

TIAViz: A Browser-based Visualization Tool for Computational Pathology Models

TL;DR

TIAViz addresses the need for flexible, browser-based visualization of digital pathology model outputs to support research and collaborative evaluation. Built as part of the TIAToolbox, it uses a Flask tile server and an SQLite-based AnnotationStore to render zoomable WSIs with layered overlays (annotations, segmentations, heatmaps, graphs) via Bokeh. The tool supports multiple overlays, graph visualizations, interactive AI workflows (e.g., HoVer-Net and potential GPT-vision integration), and enables remote or server-based demos. This open-source solution simplifies integration into Python-based workflows and enables rapid exploration of model outputs, with demonstrated applications in IGUANA, SNA, HiGGsXplore, and MesoGraph.

Abstract

Digital pathology has gained significant traction in modern healthcare systems. This shift from optical microscopes to digital imagery brings with it the potential for improved diagnosis, efficiency, and the integration of AI tools into the pathologists workflow. A critical aspect of this is visualization. Throughout the development of a machine learning (ML) model in digital pathology, it is crucial to have flexible, openly available tools to visualize models, from their outputs and predictions to the underlying annotations and images used to train or test a model. We introduce TIAViz, a Python-based visualization tool built into TIAToolbox which allows flexible, interactive, fully zoomable overlay of a wide variety of information onto whole slide images, including graphs, heatmaps, segmentations, annotations and other WSIs. The UI is browser-based, allowing use either locally, on a remote machine, or on a server to provide publicly available demos. This tool is open source and is made available at: https://github.com/TissueImageAnalytics/tiatoolbox and via pip installation (pip install tiatoolbox) and conda as part of TIAToolbox.
Paper Structure (11 sections, 1 figure)

This paper contains 11 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Technical Details
  3. Features and Capabilities
  4. Annotation and Segmentation Overlays
  5. Heatmaps and Whole Slide Overlays
  6. Graph Overlays
  7. Interactive AI Environment
  8. Other Features
  9. Example Applications
  10. Discussion and Conclusions
  11. Acknowledgements

Figures (1)

  • Figure 1: Overview of TIAViz architecture and capabilities. Panel A: High level architecture and data flow. B: Graph overlay features. C: Overlay of annotations and segmentation. D: Image-based overlays. E: Dual linked window capabilities.