Path Foundation Model
Stay organized with collections
Save and categorize content based on your preferences.
Path Foundation is a machine learning (ML) model that produces embeddings based
on digital pathology images. The embeddings can be used to efficiently build AI
models for pathology analysis-related tasks, requiring less data and less
compute than having to fully train a model without the embeddings or the
pretrained model.
Trained on large scale datasets, Path Foundation helps businesses and
institutions in healthcare and life sciences do more with less pathology data,
accelerating their ability to build AI models for pathology image analysis.
For details about how to use the model and how it was trained, see the
Path Foundation model card.
Common Use Cases
The following sections present some common use cases for the model. You're free
to pursue any use case, as long as it adheres to the
Health AI Developer Foundations terms of use.
Data-efficient classification
Path Foundation can be used for data-efficient classification tasks, including:
- Identifying tumor tissue and other distinct tissue classes
- Grading tumors
- Identifying the presence of known biomarkers to predict treatment response
- Exploration of pathology features for novel biomarker discovery
- Any variety of other feature detection and classification tasks performed on
Whole Slide Images (WSIs)
- Detecting clinical features within tissues
- Determining the type of tissue or type of stain
- Determining pathology image quality
With a small amount of labelled data, you can train a classifier model on top of
Path Foundation embeddings. Furthermore, the embedding from each tissue patch
only needs to be generated once and can be used as an input for a variety of
different classifiers, with very little additional compute.
For an example of how to use the model to train classifiers using the
CAMELYON16 Dataset, see the following
Colab notebooks:
Similar-image search
Path Foundation can also be used to find similar images within or even between
WSIs. By selecting reference patches of interest, you can use the embeddings
from Path Foundation to quantify similarity between any other set of patches
based on distances in the Path Foundation embedding space. This lets you
identify and retrieve patches most similar to your reference regions.
Next Steps
Except as otherwise noted, the content of this page is licensed under the Creative Commons Attribution 4.0 License, and code samples are licensed under the Apache 2.0 License. For details, see the Google Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.
Last updated 2025-02-11 UTC.
[[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Missing the information I need","missingTheInformationINeed","thumb-down"],["Too complicated / too many steps","tooComplicatedTooManySteps","thumb-down"],["Out of date","outOfDate","thumb-down"],["Samples / code issue","samplesCodeIssue","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-02-11 UTC."],[[["\u003cp\u003ePath Foundation is a machine learning model that generates embeddings from digital pathology images, enabling efficient AI model building for pathology analysis.\u003c/p\u003e\n"],["\u003cp\u003eIt facilitates data-efficient classification tasks like tumor identification, grading, and biomarker detection, requiring less data and compute resources.\u003c/p\u003e\n"],["\u003cp\u003ePath Foundation can be used for similar-image search within and between whole slide images, quantifying similarity based on embedding distances.\u003c/p\u003e\n"],["\u003cp\u003eBusinesses and institutions can leverage Path Foundation to accelerate AI model development for pathology image analysis.\u003c/p\u003e\n"]]],[],null,["# Path Foundation Model\n\nPath Foundation is a machine learning (ML) model that produces embeddings based\non digital pathology images. The embeddings can be used to efficiently build AI\nmodels for pathology analysis-related tasks, requiring less data and less\ncompute than having to fully train a model without the embeddings or the\npretrained model.\n\nTrained on large scale datasets, Path Foundation helps businesses and\ninstitutions in healthcare and life sciences do more with less pathology data,\naccelerating their ability to build AI models for pathology image analysis.\n\nFor details about how to use the model and how it was trained, see the\n[Path Foundation model card](/health-ai-developer-foundations/path-foundation/model-card).\n\nCommon Use Cases\n----------------\n\nThe following sections present some common use cases for the model. You're free\nto pursue any use case, as long as it adheres to the\n[Health AI Developer Foundations terms of use](/health-ai-developer-foundations/terms).\n\n### Data-efficient classification\n\nPath Foundation can be used for data-efficient classification tasks, including:\n\n- Identifying tumor tissue and other distinct tissue classes\n- Grading tumors\n- Identifying the presence of known biomarkers to predict treatment response\n- Exploration of pathology features for novel biomarker discovery\n- Any variety of other feature detection and classification tasks performed on Whole Slide Images (WSIs)\n- Detecting clinical features within tissues\n- Determining the type of tissue or type of stain\n- Determining pathology image quality\n\nWith a small amount of labelled data, you can train a classifier model on top of\nPath Foundation embeddings. Furthermore, the embedding from each tissue patch\nonly needs to be generated once and can be used as an input for a variety of\ndifferent classifiers, with very little additional compute.\n\nFor an example of how to use the model to train classifiers using the\n[CAMELYON16 Dataset](https://camelyon16.grand-challenge.org), see the following\nColab notebooks:\n\n- [Path Foundation linear classifier notebook in Colab - for Cloud Storage](https://colab.research.google.com/github/google-health/path-foundation/blob/master/notebooks/train_data_efficient_classifier_gcs.ipynb)\n- [Path Foundation linear classifier notebook in Colab - for Google Cloud\n DICOM](https://colab.research.google.com/github/google-health/path-foundation/blob/master/notebooks/train_data_efficient_classifier_dicom.ipynb)\n\n### Similar-image search\n\nPath Foundation can also be used to find similar images within or even between\nWSIs. By selecting reference patches of interest, you can use the embeddings\nfrom Path Foundation to quantify similarity between any other set of patches\nbased on distances in the Path Foundation embedding space. This lets you\nidentify and retrieve patches most similar to your reference regions.\n\nNext Steps\n----------\n\n- [Get started using the model](/health-ai-developer-foundations/path-foundation/get-started)"]]