Thats a extreme end of radio waves. such systems have several drawbacks comapared to normal radars, the main one being a small transmission range. The best systems for facial recognition will use sub mm waves. In all cases though to combine it with traditional radar systems to detect faces from 500 km will need some inegnuity and out of the box thinking. Maybe the Russians have figured out some novel method.
Quantum Radars : What Are They?
- Thread starter _Anonymous_
- Start date
Thats a extreme end of radio waves. such systems have several drawbacks comapared to normal radars, the main one being a small transmission range. The best systems for facial recognition will use sub mm waves. In all cases though to combine it with traditional radar systems to detect faces from 500 km will need some inegnuity and out of the box thinking. Maybe the Russians have figured out some novel method.
mmWaves have a fairly wide frequency range 30-100GHz plus mmWave radars are known for quite some time in military, IIRC many CIWS use them in targetting radar. I do agree that range might be an issue, which is why I said that using such a technique in weapon systems is unlikely.
Not really, deep learning systems tend to work fairly well with limited resolution.
The advent of learnt feature changed a lot many things. What was very hard with engineered feature is very much possible with learnt feature. That said you need data, getting which is linear function of money.
That paper is making a guess of low res image from the already available high res database. The issue of detecting faces realtime form a 500 km distance is something else. Its not just a issue of including better AI but also more of a fundamentally advanced hardware technology.
depends on how much stock they have in the bunkers. and what if they have connected tunnel systems inside for their supplies?
Last edited:
Few terminology clarification first. "Detection" means finding an area in an image or representation which contains an object (like a face), ANY face or face like object will be detected.
"Recognition" means matching an unseen instance (low-res or high-res) against a label (say a person's name). Label is often represented by an existing image.
So yes, it is expected that you have a high enough resolution "sample image" to match against a low resolution image seen in wild.
This is the very definition of "recognition" problem.
And yes, it is very much an issue of better "AI" because with a better "AI" you can work with much more distorted or noisy or low resolution features. Can it work from 500 KM? I don't know but radars can certainly be used to recognize faces. Especially with modern deep learnt features.
when you are worried about defining the definitions you can walk on that path alone. All the best
RADAR CANNOT DETECT FACES as per whatever is available commercially. Unless of course you rename your optical and IR detection systems to RADAR like those chinks did.
Who know where, maybe to mainland tibet, maybe into different mountains. The Chinks are there for 50 years. Who knows what they might have made there.
It could even mean the past if you go through a quantum tunnel. So you could have quantum radars capable of detecting the future in the past.
