Detect unregistered person scenario

[actran]

Let's say I can gather images of everyone who regularly walks down my driveway - my "whitelist":

• Every family members
• My mail carriers

With BI + CodeProject.AI, how can I configure to detect for unknown person event? (...meaning any person not part of my whitelist)

Some additional requirements:

• Need to work for both day and night time
• Need very low false positives, must be correct detection 99% of time

Some challenges...

• What if person is walking away versus towards camera?
• I prefer not going down face recognition route but maybe that is the only way?
• Mail carriers typically wear the same uniform every time, but my family members have different clothing. How do we handle?

 

[sebastiantombs]

You will discover that facial recognition in a PC based AI program is more of a toy than a dependable detection method. Remember, too, this is the real world and not TV or the movies.

 

[wittaj]

+1 above!!!

You will not find that with what is readily available.

Most of us have found that facial recognition is more gimmicky and novelty than anything else. If you have to put in 5 or 10 or 15 or 35 pictures or more of yourself in the system for it to recognize it is you...then you shouldn't expect much.... My success rate was under 5% so I moved on to other things LOL.

It can work in certain situations like a business that requires everyone to stop in front of the camera and the camera is at head height. Outside of that, the percentage of being accurate is probably not going to be super high. You will get a lot of false "confirmations" doing a search.

Someone here posted once how horrible it was inside his house identifying his neighbors and others as him. Another guy his kids and wife were being tagged as him.

Unless you spend the big bucks that casinos and airports have LOL.

Heck even in ideal situations like a business with the camera at ideal height and optimal lighting it fails....

What do you think of Dahua's "Uyghur Warnings"?

Good luck. Wish you no ill will. I'm not here to argue with people simply to contribute when and where I can. This is also followed by lending my views on current technology and how it impacts all of us. It's undisputed that splitting of the atom changed the world but once again man decided to...

ipcamtalk.com

 

[DanDenver]

• must be correct detection 99% of time

Not possible with a facial recognition residential install. Absolutely not possible. You would need to have each person enter a booth that has controlled lighting, controlled distance and controlled camera angle along with numerous attempts to scan the image as these AI servers will report back an unknown face within milliseconds of reporting back a known face while scanning the same person.

You may be influenced by Hollywood's overhyped representation of this technology.
Here in reality in a commercial building where people are forced to enter via a turn style, you can force the user to:
A: be in a very specific location
B: force them to slow down
C: Single file at all times no exceptions
Under these circumstances you could expect to improve your recognition percentage. But this is a commercial example, I am hard pressed to see how this could apply to a driveway or anything residential

• What if person is walking away versus towards camera?

AI that recognizes receding hair lines? This is not an option anywhere in todays world and I don't think that anyone is working on such a concern.

• I prefer not going down face recognition route but maybe that is the only way?

Is there any other option known to mankind at this time? Body gestures and clothing being attributed to a particular individual is decades or lifetimes away at best

I have used facial recognition with my DeepStack install and it is only for fun and party tricks

CodeProject.AI has the ability at this time to determine what activity an individual is engaged in. Such as "are they dancing" or "are they riding a bike"

How else can you recognize or acknowledge someone is "Jane" versus "Mary" other than their face? It is certainly not by whether they are dancing or riding a bike

• Reverse use case

You are attempting to base logic off a "reverse use case". Meaning if something is not found, do something. The challenge is that AI servers will report back unknown on a regular basis, actually alternating with known.
You would get much further by saying "if AI recognizes a face, then do XYZ". Then when an unknown is reported back, nothing happens. The user must simply continue to stand there until the AI eventually (and hopefully) recognizes them.

 

[M4DM4NZ]

You might be interested in my post here about detecting unwanted people.

 

[DanDenver]

Clarification: not unwanted people. Just people you have not logged into your system yet.

Not sure that a new phone in the neighborhood equals a person with nefarious intent.
Packet sniffing for someones probable intention - promising intellectually, but a lot of challenges to get any actionable information.

 

[M4DM4NZ]

Clarification: not unwanted people. Just people you have not logged into your system yet.

Not sure that a new phone in the neighborhood equals a person with nefarious intent.
Packet sniffing for someones probable intention - promising intellectually, but a lot of challenges to get any actionable information.

Sure, I guess you could think of it as an early warning system for possible intention, and the actionable information could be used as a trigger to start recording a camera or send an message. depends how you use the data i guess.

 

[actran]

You might be interested in my post here about detecting unwanted people.

Thanks for sharing.