FC
Francesco Castaldi
[ TRACK // PROFESSIONAL & WORK PORTFOLIO ]ID: ci-cervical-lbc

CI Cervical LBC — Deblurring & Image Restoration

Research project to clean up noise and blur from Pap smear slides. Evaluates the impact of generative diffusion models against classic mathematical deblurring to avoid medical 'hallucinations'.

PythonPyTorchDiffPIRUniversity
Cervical LBC Deblurring
FIG: Cervical LBC Deblurring

Medical Image Deblurring

Evaluates mathematical methods (Total Variation) vs Deep Learning (UNet, DiffPIR) for the restoration of LBC cytology images. Shows that while generative diffusion models produce sharp visual results, supervised UNet architectures provide the highest clinical safety against synthetic hallucinations.

[ Related Engineering Disciplines ]

Data Science & AnalyticsArtificial Intelligence & Computer Vision
[ View Upstream / Repository ]
[ ← Back to Work Portfolio ]
[ FC.SYS // ENG ]
GitHub|LinkedIn|Email
FRANCESCO CASTALDI © 2026