.

Improving Image Analysis Workflows for Large Datasets
5 February 2025
By Mark Laurie

This week, we heard from Senior Image Analysis Scientist, Hannah MacLeod, discussing improving Image Analysis for large datasets.

Hannah talked about a recent project where we developed an image analysis workflow to process a large dataset, covering tissue and cell segmentation, marker thresholding, and spatial analysis.

Along the way, we addressed several key challenges, including:

  • Refining tumour/stroma segmentation – We used Indica Labs’ HALO to train algorithms on multiple fluorescence channels. This helped optimise segmentation across different cancer types, improving accuracy and consistency.
  • Managing auto-fluorescence bleed-through – A multi-step approach allowed us to reduce unwanted fluorescence signals, improving data quality without losing critical information.
  • Ensuring effective client collaboration – Regular communication with our clients helped us adjust the workflow to their specific needs and expectations, ensuring reliable results over the course of the project.

By taking a structured and iterative approach, we improved segmentation accuracy and streamlined analysis for large datasets.

Thanks to Hannah and the team for continuing to improve our workflows and ensuring we exceed our client’s expectations at every opportunity! 👏

📩 Newsletter: Subscribe now for the latest Webinars, Videos, Articles & more! →

Find out about OracleBio's Pathology Expertise

Learn How OracleBio Can Support You

Speak to our experts

Planning your study and need advice on your image analysis? We’d love to chat!