Client wants facial recognition system that throws an alert if someone NOT RECOGNIZED is in the house. Is there such a solution?

[BrownChiLD]

I know facial recognition is somewhat good (not 100%), but for use case of DETECTING and RECOGNIZING FACES, then throwing an alert when it recognizes a face.

But this client wants the reverse, throw an alert when a face is not recognized. Use case = he wants to be alerted if a stranger is in the room.

Knowing a bit of how facial recognition works, I could imagine that most of the time the system wont recognize registered faces (negative/no recognition state) because of angle, pixelation/blurr, etc... and if this is the state that we are attaching an alert to, then the system would be throwing alerts 99% of the time, yes?

Or is there such a system already that supports this type of use case effectively?

 

[Mark_M]

Yes, Dahua 5 series NVRs have facial recognition and stranger alarm.
It requires making a database of known faces. You can easily add faces to the database when they appear, there's a 'add face' button. Or you can upload faces.
There's also the options to search for a stranger or search by a face in playback.
DMSS phone app can subscribe to the stranger alarm, or make the NVR play a sound effect file or use the relay output, or a camera with a speaker..

This is an NVR5x-I.
The '-I' or '-E/I' models have facial recognition built into the NVR. Cheaper '-4kS2' NVRs require a special facial recognition camera. E.g. NVR5208-08P-I is a 5 series, 2 HDD, 8 ch, 8 POE, with built in intelligence.

 

[MC1987]

Pretty much any Dahua recorder can do this these days

 

[wittaj]

Keep in mind that most of us have found that facial identification 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. YMMV

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 inside the house.

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

 

[BrownChiLD]

Yes, Dahua 5 series NVRs have facial recognition and stranger alarm.
It requires making a database of known faces. You can easily add faces to the database when they appear, there's a 'add face' button. Or you can upload faces.
There's also the options to search for a stranger or search by a face in playback.
DMSS phone app can subscribe to the stranger alarm, or make the NVR play a sound effect file or use the relay output, or a camera with a speaker..

This is an NVR5x-I.
The '-I' or '-E/I' models have facial recognition built into the NVR. Cheaper '-4kS2' NVRs require a special facial recognition camera. E.g. NVR5208-08P-I is a 5 series, 2 HDD, 8 ch, 8 POE, with built in intelligence.

View attachment 171251

Thanks man. Have you tried this yourself though? Is it really useful? Or more of a gimmick? because i still can;t imagine how accurate stranger detection can technically be knowing how facial recognition is basically just sampling of multiple faces in your database and for every 1 person scanned and sampled against many faces in the NVR's database, that means most of those samples will result to alarms!

Keep in mind that most of us have found that facial identification 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. YMMV

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 inside the house.

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

Thanks for the advise.

Actually, my client is ready to spend the big bucks for a proper system but I'm afraid to recommend systems, even expensive ones, because of his odd use case (Stranger Alarm) which as you have explained, many samples of many faces will result to many 'strangers' flagged before 1 is flagged recognized. But alas, I need to evaluate all options, so, hit me! Can you suggest a system used by casinos and airports that does facial recognition with stranger alarm?

 

[Mark_M]

Thanks man. Have you tried this yourself though? Is it really useful? Or more of a gimmick? because i still can;t imagine how accurate stranger detection can technically be knowing how facial recognition is basically just sampling of multiple faces in your database and for every 1 person scanned and sampled against many faces in the NVR's database, that means most of those samples will result to alarms!

The facial recognition is surprisingly accurate. Obviously depends on how good you set the camera up.
My facial recognition has always been accurate detecting the courier, family members and neighbours.
It does struggle if the person never looked towards the camera, such as side-on view. The NVR takes the best snapshot it can to identify, I have had people not recognised because of a side-on view.

Most false alerts get eliminated by setting the min/max size of a face it should detect. Similarity thresholds for known faces I keep at 80% default.
There has been occasional stranger false alerts when the facial recognition thinks a t-shirt is a face.