An end-to-end framework redacts PHI from medical images via CRNN detection and restores them with Stable Diffusion inpainting to enable privacy-preserving data sharing without losing downstream utility.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
From Redaction to Restoration: Deep Learning for Medical Image Anonymization and Reconstruction
An end-to-end framework redacts PHI from medical images via CRNN detection and restores them with Stable Diffusion inpainting to enable privacy-preserving data sharing without losing downstream utility.