Axis Communications is a network-camera company, and its people-counting story runs on that camera. Axis's own analytics have counted visitors through AXIS People Counter, which its public documentation now positions as being replaced by AXIS Object Analytics, alongside a dedicated AXIS 3D People Counter that uses stereo sensing. There are also third-party people-counting applications built on the ACAP platform that run on Axis cameras. Those are separate products from Axis's own analytics, and it is worth keeping them distinct when you compare: an ACAP app is a partner's software running on Axis hardware, not an Axis counting product.
Across all of those routes, the count is produced from a network camera. This post is for buyers weighing that against a method with no camera at all. It sets out what changes when the lens comes out of the count, and where Ariadne fits. Comparisons cite public documentation; no client relationship or disparagement intended.
What is an alternative to an Axis people counter?
An alternative to a camera-based Axis people counter is a system that records counts without capturing images. Ariadne uses Hybrid Fusion: Time-of-Flight depth sensing at entries captures geometry rather than images, and patented phone-signal sensing follows movement inside, fused in the platform. No MAC address is recorded by default, no device ID, no biometric data, and no camera is involved. The choice often comes down to whether you want a counter that runs on your camera estate, or a camera-free method with a lighter privacy-documentation footprint. Confirm any Axis people-counting detail against its current product and ACAP documentation before comparing, and verify accuracy on your own site.
To be fair to Axis: its public materials describe the counting output as numeric, produced on the camera. So the contrast to lean on is not a claim that Axis identifies people. It is simpler and more honest than that: one method still has a camera at the door, and the other has no camera at all. Everything below follows from that single difference.
Counting on a network camera vs counting without one
When the counting sensor is a network camera, you inherit the camera's status in your compliance world even when the analytics only ever emit a number. The device is optically observing people, and that is what CCTV rules, retention policies, and impact assessments key off. A camera-free counter avoids that not by processing the image more carefully but by never forming one.
Here is what the two approaches imply, side by side in plain terms:
- The camera's dual life. A network camera that counts can usually also record. That flexibility is a genuine advantage if you want counting and surveillance from one device. It is a liability if you specifically do not want a camera in that space, because the capability alone can be enough to trigger CCTV notice and works-council scrutiny.
- Retention and footage. Even where the count is the only thing stored, a camera platform has to have a clear position on whether frames are buffered or recorded, for how long, and who can access them. A depth-and-signal method forms no image, so there is no footage question to answer.
- DPIA scope. An assessment for camera-based counting has to consider image capture and the potential for identification, especially where the same hardware could be repurposed to record. A method that captures no image and no identifier keeps that scope narrower.
- Edge output does not remove the lens. Axis stating that counting runs on the camera and outputs numbers is a real privacy-by-design point in its favour, and it should be credited. It does not change that the sensor is a camera, which is the fact your signage, retention, and works-council obligations respond to.
If your reason for seeking an Axis alternative is that you do not want a camera in the counted space at all, none of the above is about Axis doing something wrong. It is about the sensing category. For the underlying framing, see what a counter records about visitors and camera-free people counting.
Axis people counting vs Ariadne, at a glance
Every Axis cell reflects its public product and ACAP documentation as published; confirm the current version on the vendor's own site, and keep Axis's own analytics distinct from third-party ACAP apps. Ariadne's column is its canonical posture. Accuracy stays a verify-on-your-own-site cell for both.
| Axis | Axis people counting | Ariadne |
|---|---|---|
| Capture method | Network camera analytics: AXIS People Counter (being replaced by AXIS Object Analytics) and AXIS 3D People Counter (stereo); third-party ACAP counting apps also run on Axis cameras | Camera-free Hybrid Fusion: Time-of-Flight depth at entries plus patented phone-signal sensing inside, fused in the platform |
| Camera / images involved | Yes, a network camera is the sensor | No camera; depth sensing captures geometry, not images |
| What it records | Axis states numeric counting output on the camera; verify current behaviour and any ACAP app separately against public documentation | No MAC address by default, no device ID, no biometric data, no image |
| Documentation burden | Camera-based: the device's optical nature typically brings CCTV signage, retention, and broader DPIA scope even where output is numeric | No image captured, so no CCTV footage to sign about or retain; narrower DPIA scope |
| Verified-on-site accuracy | Verify on your own site | Verify on your own site |
The table is a checklist, not a verdict. The row that separates the two methods in kind is the camera row. If you already run Axis cameras and want counting from them, that shared hardware is a point for Axis; if you want no camera in the count, it is the reason to look at a camera-free method.
How Ariadne counts camera-free
The camera-free claim only means something if the method genuinely forms no image, so here is exactly how it works.
Ariadne measures this with Hybrid Fusion, its patented camera-free method. Time-of-Flight depth sensing counts every visitor at the entrances, capturing geometry rather than images, while patented phone signal sensing follows movement through the interior, detecting the signals a phone emits even in airplane mode, and tracks that movement to about one-metre precision. The sensor streams both feeds to Ariadne, where Hybrid Fusion combines them into one trajectory per visit and computes counts, dwell, and paths. The streams carry no identifier: no MAC address, no device ID, no biometric data, and no camera is involved. Identifiers are stored only when a visitor explicitly opts in, which keeps the method GDPR-friendly and outside biometric territory.
Two things follow for a buyer. There is no PII captured at the sensor: the depth and signal streams carry no MAC address by default, no device ID, and no biometric data, so there is no personal data collected that would need anonymising later. And because no biometric identification or categorisation happens, the method sits outside the biometric categories the EU AI Act treats as high-risk under Annex III. For the regulatory mapping, see where counting sits under the EU AI Act; for the broader shortlist, the systems comparison and Ariadne's camera-free counting.
How to trial both fairly
A camera-based counter and a camera-free one should be judged on the same door, with the same yardstick. The yardstick is a manual ground-truth count, not a datasheet.
Choose one representative entrance, ideally busy, with the groups and lighting that expose weak counting. Run each candidate over a defined period, and for a sample of hours count the same door by hand or against a reviewed reference. Compare the system's figure to your manual figure for those hours, and you have accuracy measured on your traffic. Doing this for both the camera-based and the camera-free option, on the same entrance, turns "which is more accurate" from an argument into a measurement. To scope the test properly, see verify accuracy on site.
Cost belongs in the same exercise. If you already own Axis cameras, counting on them may look cheaper up front, while a camera-free method may change the install and privacy-documentation lines. Weigh the whole picture rather than the sticker, and if a free or trial tool is on your list, see free vs paid people counters for what each tier actually delivers.
FAQ
Do I need cameras to count people?
No. Ariadne counts with Hybrid Fusion: Time-of-Flight depth sensing plus patented phone signal sensing, never cameras. Time-of-Flight captures geometry rather than images, and signal sensing captures no MAC address by default, so the measurement involves no video, no faces, and no biometric data.
Does Axis identify people when it counts?
Axis's public documentation describes its counting as numeric output produced on the camera. So the honest contrast is not about identification; it is that Axis counts using a camera and Ariadne counts with no camera at all. Confirm the current behaviour of any Axis product, and any third-party ACAP app, against public documentation.
If Axis counts on the camera, why does the camera still matter for compliance?
Because the sensor is an optical camera, and that is what CCTV signage, retention rules, and works-council scrutiny respond to, whether or not footage is stored. Numeric-only output computed on the camera is a real privacy-by-design point, but it does not remove the lens that opens those obligations.
Can I keep my existing Axis cameras and just add Ariadne?
Ariadne is a separate camera-free system with its own hardware; it does not run on your cameras. If the goal is to count without a camera in the space, that separation is the point. If you specifically want counting from the cameras you already own, that is a case for staying on the camera-based route.
How do I compare an Axis people counter with Ariadne fairly?
Compare on what each captures, the documentation burden that follows from the sensing method, install and cost given any cameras you already own, and accuracy verified on your own site. Keep Axis's own analytics separate from third-party ACAP apps, ground every Axis claim in current public documentation, and hold both to the same manual ground-truth count on one entrance.
