Photon Statistics Based Light Classification for Optical Perceptron Development

We built a Michaelson interferometer using a 532 nm laser, this set up was able to produce interference fringes, confirming it's reliability for phase sensitive optical experiments. We will later incorporate a BBO crystal to exploit the linear electro-optic effect for phase modulation from an applied electric field--allowing for dynamic control of optical paths which is crucial for implementing perceptron like behavior. To classify light based on photon statistics, we used a single photon counting module to record photon arrival under two conditions: coherent laser light and ambient room light. The histogram of photon counts per time bin of the laser showed Poisson statistics which is expected from coherent light. In contrast, the ambient light showed superPoisson behavior, which is consistent with thermal light sources which show photon bunching effects. The statistical differences confirm our ability to distinguish & classify different types of light using photon counting--a crucial capability for optical neural network applications.