Prophesee recently announced that its event-based vision technology is designed into the new aSee glasses-EVS wearable eye tracker from 7invensun. Aimed at scientific, medical, and industrial use cases, the design features Prophesee’s GenX320 sensor that enables an eye movement sampling rate of up to 1,000 Hz.

The 7invensun wearable eye tracker features a glasses-like design and includes the Prophesee GenX320 sensor, the smallest and most power-efficient event-based vision sensor in the industry, according to Prophesee. (The GenX320 sensor was a Product of the Year Award winner in 2023 for sensors.) Leveraging the sensor’s “change perception” capability, it responds to eye movements with high sensitivity, operating only when changes are detected. Its eye-tracking sensor achieves a temporal resolution that far surpasses traditional eye trackers, enabling devices to accurately capture even the most minute eye movements, such as rapid saccade and microsaccades.

Prophesee technology strengthens 7invensun’s position in a multi-billion-dollar global market for eye-tracking technologies. (Source: Prophesee)

aSee Glasses-EVS

The launch of the aSee Glasses-EVS creates a myriad of opportunities for multidomain research in laboratory neuroscience, sports science, medical diagnosis, and cognitive processes. They can also be used in multimodal research and human factors/ergonomics research.

The aSee Glasses-EVS maintains 7invensun’s lightweight design approach. The smart glasses are reported to be comfortable to wear and not a burden to users even during prolonged use. They can also be used with contact lenses and feature detachable lenses, catering to the needs of users with different vision requirements.

Event-based vision applications are disrupting traditional solutions in applications such as autonomous vehicles, robotics, industrial automation, surveillance and security, virtual and augmented reality, biomedical, high-dynamic range imaging in satellites, IoT, healthcare, and high-speed vision.

Event-based neuromorphic vision uses sensors that mimic the human eye’s ability to focus on changes instead of capturing fixed-interval full images. Mirroring biological vision systems, the resulting benefits include lower power consumption, reduced data processing, and the ability to operate effectively in challenging lighting conditions. Event-based cameras respond quickly to changes, enabling real-time applications.

“Prophesee’s unique position provides full control of the sensor architecture, with an on-chip digital pre-processing pipeline and an on-chip custom RISC core, which allows optimization of the full data path from sensing to data transmission to application processing, and the ability to achieve unmatched energy efficiency and end-to-end latency,” said Christoph Posch, co-founder and CTO of Prophesee. “Involved techniques at the sensor level include various spatio-temporal filtering, data augmentation techniques such as dynamic ROI control, time slicing, and illuminator synchronization, targeting data rate/volume reduction and data relevance/information content optimization.”

Each pixel in the sensor acts independently, triggering an “event” when a change in light intensity exceeds a threshold. Only changes are recorded, and the resulting sparse data stream is much smaller than data generated by traditional frame-based cameras. Benefits of the technology include reduced power consumption, low latency, high dynamic range, efficient processing suitable for embedded systems, and improved motion blur.

Traditional machine-vision cameras capture images in frames per second. In comparison, event-based vision cameras are built with a neuromorphic vision sensor that captures changes in brightness rather than complete images like traditional frame-based cameras. Like human eyes and the brain, the sensors capture an image when they detect a change or event. Each pixel reports only when it senses movement, whereas in a frame-based sensor, all pixels record at the same time. In an event-based sensor, each pixel is perfectly independent; movement is captured as a continuous stream of information. Nothing is lost between frames.

The Prophesee sensor records a rotating robotic arm as a continuous stream of movement. (Source: Prophesee)

Event-based vision systems perceive the vitality of the scene while ignoring the irrelevant. They produce up to 1,000× less data than a conventional sensor and achieve a higher equivalent temporal resolution of >10,000 fps.

“All the usual strong points of event-based vision—high temporal resolution, data sparsity, and inherent low-power sensing—obviously apply perfectly to smart glasses and XR use cases,” Posch said. “This translates into more realistic, immersive, and contextually accurate user experiences, longer battery life, and lighter-weight design of devices.”

When asked if there were any surprises regarding development of the technology, Posch explained: “Notably, the use of a ‘bright pupil’ approach, using an off-axis offset/background signal with illuminator or electronically in-pixel, is a key and unique enabling technique. It came as a bit of surprise to find out that this background-signal bright-pupil technique can also be done electronically without an actual illuminator and be implemented very efficiently within each of the event pixels without increasing the size and complexity of the pixels significantly. This analog circuit-level innovation is a key ingredient for the new Prophesee eye-tracking sensor.”

The post Event-based vision sensor powers wearable eye tracker appeared first on Electronic Products.