Intelligent software for smarter video surveillance

Faceter is CCTV software with machine learning making video surveillance smart. It has real-time human face recognition and tracking, user-friendly console for all your cameras and security events accessible from everywhere: equally convenient on mobile, tablet and in browser.

Would you recommend this product?
8 Reviews5.0/5
This is really fascinating. Also, goodbye, privacy 👋🏼
Upvote (11)Share
@rrhoover Waiting for face recognition blocking startups
@rrhoover @zhansaparov it's called plastic surgery 🤣
@zhansaparov no joke, I recently heard about a startup that's working on this.
@zhansaparov @rrhoover @valereonmobile /haircut is cool, but it will not help in real situation. Hair is moving and opens different parts of face. Algorithm takes several frames from videostream to build individual vector, which will define person. I described this process with some details in the blog here: https://becominghuman.ai/how-fac...
Used to work in the industry. Casinos, jails, and hospitals would love this; especially casinos. Places like Dubai and Columbia as well. The only issue would be everyone wants onPrem and you’d have to compete with Axis. Is there still no platform for video security that supports 3rd party services as add ons yet like heroku, Salesforce, or Slack?
@kay0stheory You are absolutely right. We have recently finished POC project with one of the biggest casino chain in South Africa. They found Faceter being very useful. They had about 50% accuracy with previous solutions, and got up to 99,7% with us (LFW test result). They have black list of visitors which they have to stop at the entrance. They are not nessesary fraudsters. Sometimes they are just gambling addicted and casino can get high penalty if they accept them.
I am really intrigued by this. How many faces can it detect simultaneously? How well does it handle a crowd?
@ghundermark Thank you for your interest! In theory there is no limitation -- everything depends on hardware capacity. We tested Faceter on stadium and got 500 per sec on 1 NVidia Titan. We are going to run Faceter in decentralized infrastructure (fog computing). We will involve crypro miners into the decentralized computing network. They usually have very good hardware. So we hope to get a lot of computing power with lower cost, compared to cloud.
Can we have an opt-out hat 🎩? Or opt-out sun glasses 👓?
@valereonmobile We are working on algorithms which will recognize person by body proportions and movements. So, Faceter will identify criminal even if the one will use hat or beard, or sunglasses. Next step -- to add voice recognition. Faceter is not only a company which is working on it. This process can't be stopped already. So, let's just use these technologies for own protection. I think it is fair if individuals will have access to this kind of technologies too, not government and big corporations only.
@tchernitski imo such invasive videotaping of non consenting individuals could be criminalized under certain laws
@valereonmobile It is an advanced powerful tool which can be used for good or for bad. And it must be regulated for sure.
How is your face recognition performing? Our Docker deployable Facebox uses some state of the art tech, is more accurate than Google and Microsoft's Faces APIs, and you can train it on specific people with a single photo (we call it one shot teaching). If you're interested, please give Facebox a try! (http://machinebox.io)
@machineboxio We reached 99.78% by LFW benchmark, using just one NN, which we really use in our product.