Anotherrelatedworkarethedepth-basedextensionstoSVG 2 proposedbyJeanLeFeuvre[12]. Theseextensionspermitthatan SVG application be executed on multi-view displays. Le Feuvre also proposes a graphics architecture for supporting such applications and discusses two implementations: one based on GPU, developed mainly for testing purposes, and another one based on FPGA (Field Programming Gate Array) targeted at low-cost platforms. Both solutions are integrated into the GPAC multimedia framework 3. The second solution is the one closest to this work. It produces a 2D+Z representation of the entire scene and sends it to FPGA. The FPGA is then responsible for generating the views required by 3D displays using DIBR. In comparison with Le Feuvre’s work, the graphics composition technique presented in this paperisimplementedinOpenCLandachievesbetterrenderingresults. This occurs because our proposal keeps as much occlusion information as possible (in the LDI structure) while compositing the different media objects.