Object classification through a single-pixel detector
1–10 of 13 posts
Re: Object classification through a single-pixel detector
#2Apparently the idea here is to substitute a bunch of convolution kernels with, well...optics
Re: Object classification through a single-pixel detector
#3Apparently the idea here is to substitute a bunch of convolution kernels with, well...optics
That would simplify the electronics for applications where electronics need to be radiation hardened.
Re: Object classification through a single-pixel detector
#4Any idea if they used any type of nonlinearities in the optics? It's hard to see this being powerful without nonlinearities.
Re: Object classification through a single-pixel detector
#5Takeaway from roughly skimming the article: the single pixel detector is a plasmonic nanoantenna, which seems to be tunable. This is placed behind a set of optical diffraction gratings that have been optimized to the task using machine learning and essentially transform the spatially encoded information in the input into a spectral encoding. Both the gratings to encode the information and the decoder have been designed using machine learning. There is no dynamic illumination and the imaged object is static in front of the optics. All of this was done in the THz range.
Re: Object classification through a single-pixel detector
#6Takeaway from roughly skimming the article: the single pixel detector is a plasmonic nanoantenna, which seems to be tunable. This is placed behind a set of optical diffraction gratings that have been optimized to the task using machine learning and essentially transform the spatially encoded information in the input into a spectral encoding. Both the gratings to encode the information and the decoder have been design…
Presumably you could invert it as well and generate images from a single ‘color’. Create a passive spectroscope that converts the observed spectrum to text identiying the chemical compounds it represents.
Re: Object classification through a single-pixel detector
#7Apparently the idea here is to substitute a bunch of convolution kernels with, well...optics
That would simplify the electronics for applications where electronics need to be radiation hardened.
I'm not so sure, this uses electronics operating in THz range, and the crystals used are also likely vulnerable.
The main gain seems to be energy use and size, thing is very efficient.
Re: Object classification through a single-pixel detector
#8My understanding is that it generates parallel to the rgb information, a depth-map and segments encoded also in specific spectrum of colors. This removes the burden of heavy image processing (edge-detection, etc) which improves performance. The segments can be trained for classification.
Re: Object classification through a single-pixel detector
#9Takeaway from roughly skimming the article: the single pixel detector is a plasmonic nanoantenna, which seems to be tunable. This is placed behind a set of optical diffraction gratings that have been optimized to the task using machine learning and essentially transform the spatially encoded information in the input into a spectral encoding. Both the gratings to encode the information and the decoder have been design…
Presumably you could invert it as well and generate images from a single ‘color’. Create a passive spectroscope that converts the observed spectrum to text identiying the chemical compounds it represents.
Hard-coded "dreaming" of a an optical "net".
Re: Object classification through a single-pixel detector
#10that's racist