Disclaimer: I work at Density (https://www.density.io) building an anonymous people counter.
MAC address tracking is a good simple way to get an approximate number of people: it's very easy to install, it requires only a WiFi antenna, and the data is easy to translate into a count. However, there are privacy and legal concerns which have prompted phone vendors to obfuscate this data. And there can easily be zero (or more than one) broadcasted MAC address per human, even when filtering by OUI.
Retailers have been using specialized "people count" technology like infrared break-beams, thermal cameras, and CCTV+CV systems for a long time. CCTV is the most accurate by far, but it's also commonly considered to be an invasive level of surveillance (and rightly so). It's also not particularly accurate in adverse situations. At Density, we looked at (and tried) many of these technologies - but ultimately found them lacking because they were either too invasive or too inaccurate. The device we've built uses a lower-resolution depth-only sensor because it can return extremely accurate results, without having the capability for facial recognition or other analysis techniques that are harmful to privacy. So far the technology is working very well - with an algorithm based on a deep-learning "human classifier" we're seeing accuracy above 99% in many of our deployments.
Here's a cool (and in-progress, excuse our dust) summary of some issues that the technology has to navigate: https://faq.density.io/algo/
The concerns you raise are true in corner cases, but the general use case should be good enough. Your solution also seems to require mounting one of your devices at the entryway of every room for which you wish to monitor occupancy. Seems like it could be costly, especially if doing so requires wiring power up to it.
That is the other side of our approach, yes. Active depth sensing and on-device deep learning means non-trivial power and speed requirements, so the hardware isn't as cheap as we'd like it to be. I'm not up to date with the latest developments on the hardware team, but they're working aggressively on our BOM and we're hoping to take advantage of cheaper and lower-power SIMD/GPU chips in the future. Our current-gen model works with Power over Ethernet, and right now we're most often selling to customers with large/valuable real-estate portfolios because they can get a higher ROI from accurate count data.
We've switched from a mobile phone MAC address based location analytics people counter system(I believe founded by someone who worked on Google Analytics) to a stereo camera based system(bellwether).
We switched as our original provider's business model shifted(I believe in part due to the change in MAC addressing by telecom providers?).
We believe strongly in personal freedom and personal data security.
But we are scratching our head as far as what we can do to ethically and affordably better understand and serve our customers in an integrated(digital & meatspace) way.
I'm not sure if our product would be able to meet your needs right away, but reach out to sales@density.io if you'd like more information about what we have.
Without knowing what specific depth sensor you're using, most of them are essentially B/W IR cameras so it seems a bit disingenuous to say that is better for privacy.
The depth map you provide doesn't contain the full output of the sensor, but how is that any different than using cameras but not passing the data up any higher?
If you look at this video:
https://www.youtube.com/watch?v=YOKMx7EDVys you can see the kind of image that Density has access to, though their depth map they show is only what's on the left side.
No part of our system uses amplitude data, not even on the local device. We only rely on phase offset/depth data. However you're correct that simply choosing a particular technology cannot automatically "solve" privacy, and it has to remain a priority indefinitely.
Looks cool!
I’m wondering how you would deal with cumulative errors. Lets say you have a meeting room with one door with one sensor, and the sensors has a 99% accuracy. If 100+ people walk through the door, the sensor could still report 1 or 2 persona in meeting room, while nobody is there.
Do you have work arounds for this problem?
Yes, we have active R&D projects focused on addressing this kind of drift. In smaller rooms where the environment is favorable, we will usually see the device count for a whole day without any mistakes. But in general there are diminishing returns with that strategy, and there are lots of possible ways to correct drift error. I'm not sure how much more I can say right now.
I work on the software side, so I can't really comment with authority, but this looks like it could be an interesting lower-frequency/lower-power approach to active sensing. From the preview site, it seems like the "resolution" of this model may be too low to separate and count a crowd of people. However I'll forward this on to the hardware team just in case, thanks!
Isn't Soli more for natural user interface (NUI) detection via radar wave disruptions? I'm not sure it would provide very granular detection and he'd have a lot of false positives.
I was thinking more in terms of lidar tracking of people. It's been around for nearly a decade and works quite well.
Right, Soli is built for NUI. However, it does seem like it might fit this use case as well:
Distance - I'd imagine that the distance from the top of a door frame to the top of the head of the average shopper is roughly equivalent to the distance from the Soli device to someone's hands.
Software - The Soli hardware streams radar contacts to software trained to separate and classify them, not too unlike the approach that Density is already using.
Benefits over a CV approach might come in power usage and lower complexity in the contact classification software.
Unfortunately, as far as I know LIDAR is still prohibitively expensive for this. Someone did invent a solid-state LIDAR system recently, which seems like a good sign.
Fair point. I was solely thinking from a technical solution not the practicality around deployment. My employer has biased my bubble on thought process. "Oh, yes, this is easy to solve, and I see it on campus (Disney - imagineering office).
Really interesting. What sort of limits/constraints do you have? For example, I've seen older methods have trouble maintaining tracking accurately at 12-24+ people in an room size area.
Since the sensor has a certain operating power, it does not work as well on very high doorways (>10ft) or in direct sunlight. Also the field of view cannot cover a very wide doorway (>8ft). Obstructions, reflections, and odd room geometry are common problems, but they can often be solved in software/configuration.
Within spec, the algorithm does well tracking 8+ people at a time, and even when a line of people extends through a doorway we usually don't have to track more than that.
The hard thing about that approach might be handling conflicting signals and determining a "source of truth" in adverse situations. Generally speaking, there are variables in the real world that will trip up any given strategy (even if you were to station a human at each doorway with a hand counter). And while nothing can count perfectly 100% of the time, the difficulty is in finding an affordable strategy that will be really close or perfect most of the time.
Here's a neat demo of a multi-sensing device (not ours) which combines many signals to guess the activity taking place in a room: https://www.youtube.com/watch?v=aqbKrrru2co
guscost correctly describes the limitations of MAC address tracking. In my experience building Aura Vision [1], we've also discovered MAC tracking is no longer GDPR compliant, because an identifier about a person is stored indefinitely. This gives retailers the ability to measure the same identifier returning. The same goes for Bluetooth/BLE tracking.
We are a CCTV/deep learning system that uses existing infrastructure (think old school grainy security cameras), and we're are also able to capture additional information like age and gender of a person. Unlike other invasive/HD CCTV systems, we don't use facial recognition, and we also work over very wide areas, not just over restricted doorways.
MAC address tracking is a good simple way to get an approximate number of people: it's very easy to install, it requires only a WiFi antenna, and the data is easy to translate into a count. However, there are privacy and legal concerns which have prompted phone vendors to obfuscate this data. And there can easily be zero (or more than one) broadcasted MAC address per human, even when filtering by OUI.
Retailers have been using specialized "people count" technology like infrared break-beams, thermal cameras, and CCTV+CV systems for a long time. CCTV is the most accurate by far, but it's also commonly considered to be an invasive level of surveillance (and rightly so). It's also not particularly accurate in adverse situations. At Density, we looked at (and tried) many of these technologies - but ultimately found them lacking because they were either too invasive or too inaccurate. The device we've built uses a lower-resolution depth-only sensor because it can return extremely accurate results, without having the capability for facial recognition or other analysis techniques that are harmful to privacy. So far the technology is working very well - with an algorithm based on a deep-learning "human classifier" we're seeing accuracy above 99% in many of our deployments.
Here's a cool (and in-progress, excuse our dust) summary of some issues that the technology has to navigate: https://faq.density.io/algo/