On Learning from Ghost Imaging without Imaging

  • 2019-03-14 14:04:51
  • Issei Sato
  • 3


Computational ghost imaging is an imaging technique with which an object isimaged from light collected using a single-pixel detector with no spatialresolution. Recently, ghost cytometry is proposed for an ultrafastcell-classification method that involves ghost imaging and machine learning inflow cytometry. Ghost cytometry skipped the reconstruction of cell images fromsignals and directly use signals for cell-classification because thisreconstruction is the bottleneck in a high-speed analysis. In this paper, weprovide a theoretical analysis for learning from ghost imaging without imaging.


Introduction (beta)



Conclusion (beta)