ankile/Adversarial-Diffusion

Code for a paper exploring using diffusion models to defend neural networks against adversarial attacks

29
/ 100
Experimental

This project helps medical professionals, particularly those working with histopathological images, ensure the reliability of AI models used for diagnosing conditions like cancer. It takes potentially compromised tissue scan images as input and processes them to remove malicious alterations, producing cleaned images that AI models can classify more accurately and reliably. This makes AI-assisted diagnostics more trustworthy for pathologists and oncologists.

No commits in the last 6 months.

Use this if you are deploying AI models for medical image analysis and need to protect them from subtle, malicious changes that could lead to incorrect diagnoses.

Not ideal if your primary concern is improving the baseline accuracy of an AI model rather than its resilience against adversarial manipulation.

medical-imaging histopathology AI-safety cancer-diagnostics image-purification
Stale 6m No Package No Dependents
Maintenance 0 / 25
Adoption 5 / 25
Maturity 16 / 25
Community 8 / 25

How are scores calculated?

Stars

9

Forks

1

Language

Jupyter Notebook

License

MIT

Last pushed

Jan 12, 2024

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/diffusion/ankile/Adversarial-Diffusion"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.