In photogrammetry-based crime scene reconstruction, what is the minimum recommended number of photographs of a single object from different positions to generate an accurate 3D model?

Remaining questions and full explanations are VIP-exclusive.