Shop Hybrid cap w/ Tango + mp4->VisualSFM

sketchfab

Human: A hybrid model created using principal point cloud captures with Lenovo Phab2 Pro and the Matterport Tango app. These were exported and aligned in meshlab, where originally transferred vertex colors to the reconstructed mesh and then parameterized 'trivial - per triangle'. The result looked remarkably good without actual texture, but had pinline shadow artifacts around most of the faces. As a follow-up, I made a 2:00 minute UHD video capture on a vanilla Samsung S6. Then ran FFMPEG frame capture at 2fps and fed all images through VisualSFM. This took about 30 minutes of compute time. I used Cloud Compare's scaling-alignment tool (4-point picking) to scale the Tango Mesh to match the scale of the VisualSFM model, with the hope that raster registration could then be applied to a better mesh produced from the tango capture. I only captured video on a subset of the overall tango scan, those outer sections are black.