Abstract
We consider sampling system optimization in the context of deep-learning-based data fusion for dynamic high depth of field events, such as baseball. We show that multiscale spatio-spectral-focal-temporal sampling is ideal.
Year