Diffractive optical network for seeing through diffusive or scattering media

A method of seeing through diffusive and/or scattering media to capture an accurate optical image, optical signal, or optical data comprising:

• passing an input optical images, optical signals, or optical data through an unknown diffusive and/or scattering media and into a diffractive optical network comprising one or more optically transmissive or reflective substrate layers arranged along an optical path, each of the optically transmissive or reflective substrate layer(s) comprising a plurality of physical features formed on or within the one or more optically transmissive or reflective substrate layers, and having different transmissive or reflection properties as a function of local coordinates across each substrate layer, wherein the one or more optically transmissive or reflective substrate layers and the plurality of physical features collectively define a trained mapping function between (1) an input optical image, input optical signal, or input data to the diffractive optical network and (2) an output optical image, output optical signal, or output optical data from the one or more optically transmissive or reflective substrate layers, the trained mapping function reconstructing the optical image, optical signal, or optical data that passes through the unknown diffusive and/or scattering media;
• and
• projecting the output optical image, output optical signal, or output optical data onto an observation plane and/or one or more optical detectors configured to see or capture the output optical image, output optical signal, or output optical data resulting from the diffractive optical network.

A method of forming a diffractive optical network that sees through diffusive and/or scattering media to capture accurate optical images, optical signals, or optical data comprising:

• training with at least one computing device a diffractive optical network model formed by one or more substrate layers to reconstruct an optical image, optical signal, or optical data that passes through the diffusive and/or scattering media that would otherwise degrade the optical image, optical signal, or optical data, the diffractive optical network model having a plurality of optically transmissive and/or optically reflective physical features located in different locations in each of the one or more substrate layers of the network, wherein the training comprises feeding an input plane of the diffractive optical network model with training images, training optical signals, or training optical data along with a plurality of different diffusive and/or scattering media, followed by computing an optical output of the diffractive optical network model through optical transmission and/or reflection resulting from the diffractive optical network and iteratively adjusting transmission and/or reflection coefficients for each substrate layer of the diffractive optical network until optimized transmission/reflection coefficients are obtained to reconstruct the optical image, optical signal, or optical data that passes through a new diffusive and/or scattering media;
• and
• manufacturing or having manufactured or otherwise fabricated a physical embodiment of the diffractive optical network comprising one or more transmissive and/or reflective substrate layers having physical features that match the optimized transmission/reflection coefficients obtained by the training.