Object voxeliztion by filtering
Miloš Šrámek, Arie E. Kaufman · 1998
Bk&ze dL~erent@ers uedfor the vozltition of ana~fial~ descr&ed objects.flk show ~when the VOX1 model is usedfor visd-adom theflter design is refatedto the subseguentrendering phe, mely Zhe gradient estimation lechn~m Our theoredcal and epertinti ana~ses show tha~in order 10 avoti a systematic error in normal estimatio~the density profi near the su~ace shod be Iinear~proportion to the dktance @m the su~m~ Oprimd ~hiclmess of this transient regwn has been estimated to be about3 VO.W1 units.Based on these results, we propose a techn~ue .forvo.witiation ofarbitrqparametric s@aces, using a hierarchical subdivision of rhe 2D s~~e domati CR Categories 1.35