This explains how to make a regressive network from a few simple examples. By taking what one might call a polyhedron of angles of the object, the source picture behaves as a thesaurus of the standing class of the object. Thus, it should be easily recalled by a TensorFlow AI.
https://www.dropbox.com/s/wa6chu5qs97a49a/ad-pulse-ai_v3.pdf?dl=0
This explains how to make a regressive network from a few simple examples. By taking what one might call a polyhedron of angles of the object, the source picture behaves as a thesaurus of the standing class of the object. Thus, it should be easily recalled by a TensorFlow AI.
This is my self-written AI. I just don't have access to the pictures like in need in the paper. https://github.com/wise-penny/LIDSx