SVIA combines a semantic segmenter, a latent diffusion inpainter, and a harmonizer to anonymize street view images by replacing people, vehicles, signs, roads, and buildings with plausible synthetic content.
LDFA: Latent diffusion face anonymization for self-driving applications,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
SVIA: A Street View Image Anonymization Framework for Self-Driving Applications
SVIA combines a semantic segmenter, a latent diffusion inpainter, and a harmonizer to anonymize street view images by replacing people, vehicles, signs, roads, and buildings with plausible synthetic content.