We address this limitation by formulating a multi-level architecture that is end-to-end trainable. Due to memory limitations in current hardware, previous approaches tend to take low resolution images as input to cover large spatial context, and produce less precise (or low resolution) 3D estimates as a result. We argue that this limitation stems primarily form two conflicting requirements accurate predictions require large context, but precise predictions require high resolution. Although current approaches have demonstrated the potential in real world settings, they still fail to produce reconstructions with the level of detail often present in the input images. Recent advances in image-based 3D human shape estimation have been driven by the significant improvement in representation power afforded by deep neural networks.
0 kommentar(er)
