Transform the subject from the input image into appealing claymation character, keep likeness. Handmade plasticine style, ...
Abstract: Recent open-world representation learning approaches have leveraged CLIP to enable zero-shot 3D object recognition. However, performance on real point clouds with occlusions still falls ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results