Law enforcement recorders, handheld computers, cloth control balls, smart helmets AI face recognition.
Hello everyone!I'm Zhao Gong, and today we're going to talk about AI recognition capabilities.
Like this law enforcement instrument of ours.
And this handset.
There's also the ball in control.
All support this AI recognition function.
The AI recognition function.
It is mainly divided into front-end recognition and back-end recognition.
In this issue, we will focus on front-end identification.
The so-called front-end identification.
It is directly identified by the front-end device.
his entire database.
It's in the memory of the front-end device.
He doesn't need to go through the network.
You can base the database inside the fuselage memory.
Perform comparative face recognition.
AI recognition functions such as license plate recognition.
The benefits of it are to say.
In the absence of an internet connection.
He can also do this AI recognition.
This will be a little bit lower in latency.
Relative back-end recognition.
One mechanism for back-end identification is:
Equivalent to yes. The front-end equipment first streams the sound.
went to give a back-end one.
One device for edge computing.
The data is then processed by the back-end device.
So this advantage of front-end equipment is:
It doesn't require a network.
It can be quickly based on this device inside.
In-memory database.
Then quickly identify it.
Like in this mall.
There are some people who have joined the blacklist.
Or a mountain resort.
He was not allowed to enter.
Take a picture of the faces of these people.
After entering our front-end device.
Once these people.
When you want to enter this resort.
He'll call the police.
The front desk will call the police.
There will also be an alarm in the background.
This way. Our staff.
You can expel him quickly.
Don't let him in.
Such a feature.
But his shortcomings are also obvious.
That is, his database cannot be particularly large.
Like the general word.
Big database words.
That's a lot. For example, tens of millions of faces.
A million faces. Words like this one.
He's a big database that's handled through the backend.
Carry out back-end batch processing.
Let's talk about front-end processing in a nutshell.
It is applied to some small databases.
Then there is no need for a network.
Quickly help you identify.
After the recognition.
Problematic. He then went through the network.
Go upload it to your server.
So that you have a backup.
What about the usefulness of back-end processing?
It may be to search a larger database.
Then to determine whether it is a criminal or not.
Or some other behavior.
But this needs to be done through the Internet.
There is a certain delay.
That this issue. Then let's focus on it.
Some of the features and demos identified by the front-end.
Now let me show you.
This body camera.
This feature of front-end face recognition.
This is front-end face recognition.
He will call the police.
The character will then be recognized.
Then upload to this backend.
Then on the backstage.
You can see a display of this platform.
Here's a demonstration.
Body worn cameras.
An effect of front-end license plate recognition.
All with voice alarms.
There will also be corresponding records in the background.
Let's take a look.
This face recognition is in the background.
In other words. We can put the front-end of this database in advance.
Refine it. Let's put some faces in.
After the entry.
For example, this vehicle identification.
Face recognition. It's all about entering it in advance.
For example, your license plate number.
And the human face. After entering in advance.
Synchronization on the device side.
That is to say.
This database will be ** into the full device.
And then the front-end device.
It's good to recognize the faces you typed.
License plates are also good. Later.
will take the initiative to upload this data to the background.
Then we will have the corresponding data in the background.
Let's take a look.
That's some of the data.
There will be a ** of this capture.
Point in time, then address.
His name. Avatar then similarity.
And this person who collected it.
All of them are. **To Shield Soul Technology: