Lossless contour coding using elastic curves in multiview video plus depth

摘要: Multi-view video plus depth is emerging as the most flexible format for 3D representation, witnessed by current standardization efforts ISO and ITU. In particular, in arguably important information lies object contours. As a consequence, an interesting approach consists performing lossless coding of contours, possibly followed lossy per-object values. this context, we propose new technique based on elastic deformation curves. Using square-root velocity representation elements space curves, can model continuous evolution deformations between two reference contour An elastically deformed version contours be sent to decoder with reduced cost used side improve actual contour. Experimental results several multiview sequences show remarkable gains respect techniques state art.