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.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 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.