Datensatz

Zusammenfassung

We present a technique for accurate denoising of time-varying
range data. It is inspired by the idea of similarity-based
non-local image filtering and spatio-temporal bilateral filtering
for video processing. We build upon both ideas and are to our
knowledge the first method which extends them to time-varying
geometric data. Our proposed algorithm is easy to implement,
preserves fine shape features and produces an accurate and
homogeneous smoothing result in the spatial and along the time
domains.