mahmoodlab/CLAM
CLAM is a deep-learning pipeline for data-efficient whole slide image classification in computational pathology using weakly-supervised attention-based multiple instance learning.
Not currently ranked — collecting fresh signals.
star history
Clustering-constrained Attention Multiple Instance Learning (CLAM) is a method for classifying whole slide histopathology images using only slide-level labels without requiring region-of-interest annotations. The pipeline includes segmentation, patching, feature extraction, and attention-based training with instance-level clustering to identify diagnostically relevant sub-regions. It has been validated across multiple datasets including TCGA data and smartphone microscopy images.
Frequently asked
- What is mahmoodlab/CLAM?
- CLAM is a deep-learning pipeline for data-efficient whole slide image classification in computational pathology using weakly-supervised attention-based multiple instance learning.
- Is CLAM open source?
- Yes — mahmoodlab/CLAM is open source, released under the GPL-3.0 license.
- What language is CLAM written in?
- mahmoodlab/CLAM is primarily written in Python.
- How popular is CLAM?
- mahmoodlab/CLAM has 1.7k stars on GitHub.
- Where can I find CLAM?
- mahmoodlab/CLAM is on GitHub at https://github.com/mahmoodlab/CLAM.