Hi! Currently I’m experimenting with RecFusion to build a setup that doesn’t need a turn table. A kind of “one shot” solution. The scanning should not take more than 10 seconds. Has anyone already made experience with such a setup?
My current setup consists of a cage of 15 Intel RealSense 435i. This is a 5 corner with 3 cameras per corner. The cameras are located at the height of 90cm, 170cm, 210cm and are about 80-90cm away from the person to be scanned.
The results looks like this:
The model itself looks okay, except for the face. I would like to somehow improve the texture, as well as the level of detail of the face. Does anyone have any tips on how to improve this setup? What does an optimal setup look like anyway?
Thank you!
