The newly introduced always-on visual sensing system is built on a low-power architecture that operates independently of the conventional image capture system. By integrating AI vision recognition models and tiered visual wake-up technology, the system enables continuous visual sensing without requiring the host system to remain active. At the core of this architecture is Himax’s WiseEye AI processor, which leverages the Arm Cortex-M55 and Ethos-U55 AI acceleration architecture to enable real-time on-device visual AI inference and human and event detection, with industry-leading ultralow power consumption of just 5 milliwatts (mW). When it detects relevant events or human presence, the processor selectively wakes the host system for further AI processing, minimizing unnecessary activity. This significantly reduces overall system power consumption and extends the battery life of smart glasses.
With the rapid advancement of AI agents and multimodal large language models, AI applications are evolving from passively responding to user commands toward continuously sensing their surroundings, understanding real-time context, and proactively delivering services. Worn at eye level and offering a first-person perspective, smart glasses can continuously capture information about the wearer’s surroundings, making them an important device for next-generation AI agent applications.
However, maintaining continuous visual sensing on a limited battery and in a lightweight form factor has long been a major challenge for the industry. To address this, Himax’s WiseEye combines ultralow power always-on sensing with endpoint AI, processing visual data on the device in real time and waking the host system only when needed. This approach reduces overall system power consumption while maintaining always-on visual AI sensing of the surrounding environment, laying a critical foundation for future AI agent applications. It also enables smart glasses to evolve toward "proactive memory, proactive services, and proactive decision-making," positioning them as personal AI assistants that understand user context and provide real-time support.