KNT applies key-conditioned nonlinear obfuscation to split-inference features, cutting re-identification AUC from 0.635 to 0.586 with 0.15 ms overhead and under 1 pp accuracy loss.
Medical imaging deep learning with differential privacy.Scientific Reports, 11:13524
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2026 2verdicts
UNVERDICTED 2representative citing papers
Evaluating DPSGD clipping methods on medical segmentation shows prior assumptions fail in this domain, but adding morphological refinement and an adaptive DP-Morph variant improves utility under privacy constraints.
citing papers explorer
-
Keyed Nonlinear Transform: Lightweight Privacy-Enhancing Feature Sharing for Medical Image Analysis
KNT applies key-conditioned nonlinear obfuscation to split-inference features, cutting re-identification AUC from 0.635 to 0.586 with 0.15 ms overhead and under 1 pp accuracy loss.
-
From Gradient Clipping to Structural Refinement: Improving DPSGD for Medical Image Segmentation
Evaluating DPSGD clipping methods on medical segmentation shows prior assumptions fail in this domain, but adding morphological refinement and an adaptive DP-Morph variant improves utility under privacy constraints.