3 ms·
Run the algo on any reflective objects in the field of view, feed the outputted audio to a transcription NN. Save all high confidence transcribed voice as compr
by sbierwagen 8y ago
Run the algo on any reflective objects in the field of view, feed the outputted audio to a transcription NN. Save all high confidence transcribed voice as compressed txt. Save the source video if you pass a threshold of danger keywords. ("bomb", "kill", "prime minister", "didn't pay my tv licence")
- jdietrich 8y agoBest case scenario, you'll need a nuclear power station and a few dozen data centers full of ASICs. Oh, and an internet's worth of extra bandwidth.