‘Can see through walls’: Open-source WiFi tool detects human movement behind walls without cameras

A newly released open-source software project is drawing global attention for demonstrating how ordinary WiFi signals can be used to detect human movement behind walls without cameras. The project, known as WiFi DensePose, analyses subtle changes in wireless signals to estimate a digital skeleton-like representation of people moving inside nearby rooms. In demonstrations shared online, the system was able to track body posture, movement and potentially breathing-related motion through walls in real time. While the concept may sound futuristic, researchers say it builds on more than a decade of scientific work showing that radio signals can act as a powerful sensing tool capable of detecting human presence without relying on cameras or wearable devices.

The viral project was developed by a programmer known online as ruvnet, who released the code publicly under an open-source licence on GitHub. The system combines WiFi signal sensing, advanced signal processing and machine learning models to translate radio signal reflections into estimates of human body movement.

Although the recent implementation has gained attention for its accessibility, the underlying science has been studied for years.

Researchers at institutions including Carnegie Mellon University have previously demonstrated that WiFi signals can be used to estimate human body poses.

One influential study, titled DensePose from WiFi, showed that neural networks could map WiFi signal patterns to detailed human pose estimates. According to the researchers, radio signals interact with the human body in predictable ways, allowing algorithms to infer posture and movement from the changes in those signals.

As Dina Katabi, whose earlier research pioneered WiFi-based motion detection, explained in earlier work on radio sensing:

“Wireless signals don’t just carry data. They also carry information about the environment they move through.”

WiFi routers constantly emit radio waves in the 2.4 GHz and 5 GHz frequency bands. These waves spread through a room, bouncing off walls, furniture and human bodies.

The key information used by the software is something known as Channel State Information (CSI). CSI describes how the strength and phase of WiFi signals change as they travel between transmitters and receivers.

When a person moves, even slightly, those signal patterns shift. Advanced algorithms can analyse these tiny fluctuations to identify motion and estimate where parts of the body are located.

The system processes the CSI data using neural networks that generate outputs conceptually similar to motion-capture systems used in animation or sports analysis. The result is a skeletal outline representing the position and movement of a person inside a room.

In some experiments, the software has also been able to detect micro-movements of the chest, allowing it to detect subtle breathing-related movement.

Despite the viral claims circulating online, the system cannot currently be activated on most ordinary household routers.

To work properly, the technology requires specialised hardware that can capture raw WiFi signal data. Many demonstrations use multiple ESP32-S3 microcontroller boards fitted with external antennas.

These small devices act as dedicated sensors that collect CSI data from WiFi transmissions. The data is then sent to a local computer where the software processes it using programs written in Rust or Python.

The entire system can run locally without relying on cloud computing, meaning the analysis happens directly on the device.

This approach allows developers to experiment with WiFi-based sensing without expensive laboratory equipment.

Scientists have long been exploring ways to use radio signals as an alternative to cameras or radar for detecting movement.

Radio waves have several advantages:

Because of these properties, WiFi sensing has attracted interest for several applications.

Researchers believe the technology could be useful for healthcare monitoring, such as detecting breathing patterns or identifying when elderly patients fall at home.

Another potential application is search and rescue operations, where radio signals could help locate survivors trapped under rubble after earthquakes or building collapses.

At the same time, the idea of using WiFi signals to detect movement behind walls has sparked debate about surveillance and privacy.

Some experts warn that if the technology becomes widely accessible, it could theoretically be misused to monitor movement without people’s knowledge.

According to Serge Egelman, technologies that transform everyday infrastructure into sensing systems can raise new ethical questions.

“When devices designed for communication suddenly become sensors, it changes the privacy landscape,” Egelman said in discussions about emerging sensing technologies.

Supporters argue that WiFi sensing systems also have advantages over cameras because they do not record identifiable visual images.

Instead, they generate abstract motion data rather than photographs or video footage.

For many researchers, the open-source project represents a glimpse into a future where wireless networks do more than simply connect devices to the internet.

Scientists increasingly view radio signals as another form of environmental sensing, similar to radar or lidar.

As wireless technology becomes more powerful and machine learning algorithms improve, everyday infrastructure such as WiFi networks could potentially be used to detect movement or support health-monitoring systems in the future.

The open-source release of WiFi DensePose shows how rapidly these ideas are moving from academic laboratories into the hands of developers and hobbyists.

Whether the technology ultimately becomes a tool for healthcare and safety or raises new concerns about privacy may depend on how society chooses to regulate and use it.