Page 15 - Demo
P. 15
The reconstruction does not benefit from additional data beyond a minimal threshold
• Required data ensemble for best results is ~10 shots (compared to 39 MCP channels)
• Once that threshold is met, reconstruction has “learned” the source distribution completely
• Deterministic calibration method needs 70+ shots to achieve similar results (~2 / channel to smooth out experimental errors)
Presenter: Gabriel Player
15
