HyperAIHyperAI

Command Palette

Search for a command to run...

Ubicept Expands Perception Technology for Reliable Autonomous Systems Using Standard Cameras

Ubicept, a computer vision startup founded by experts from MIT and the University of Wisconsin-Madison, has announced the release of the Ubicept Toolkit, a new software platform designed to enhance visual data quality and reliability for machines. The toolkit brings the company’s physics-based imaging technology to both next-generation single-photon avalanche diode (SPAD) sensors and conventional CMOS cameras, offering a unified solution for a wide range of applications in robotics, automotive, industrial sensing, and AR/VR. The core of Ubicept’s innovation lies in its physics-based approach to image processing, which goes beyond traditional AI-driven video enhancement. Unlike models trained on data that may not reflect real-world physical behavior, Ubicept’s system models the actual physics of light and motion, enabling it to deliver more accurate, trustworthy, and high-fidelity visual data—especially in extreme conditions like low light, high dynamic range, and fast motion. This is particularly critical for autonomous systems that rely on precise perception to make real-time decisions. The Ubicept Toolkit is designed to work with any modern vision system that outputs raw, uncompressed video, including widely available CMOS cameras. This means developers can immediately improve image quality and reliability without replacing existing hardware. For more advanced applications, the toolkit is also optimized for SPAD sensors—such as Pi Imaging Technology’s SPAD Alpha, a 1-megapixel single-photon camera that offers zero read noise, nanosecond-level exposure control, and frame rates up to 73,000 fps. The SPAD Alpha is ideal for scientific, medical, and high-precision industrial applications, and the Ubicept Toolkit enhances its raw photon data into clear, usable imagery for computer vision tasks. The toolkit includes the Ubicept Photon Fusion (UPF) algorithm for real-time video processing, as well as FLARE (Flexible Light Acquisition and Representation Engine) firmware to optimize photon capture. It also comes with white-glove support for early adopters, ensuring a seamless integration into existing development workflows. Sebastian Bauer, CEO of Ubicept, emphasized the technology’s transformative potential: “Our physics-based approach captures the full complexity of motion, even in low-light or high-dynamic-range conditions, providing more trustworthy data than AI-based video enhancement.” Tristan Swedish, CTO, added that the goal is to turn raw photon data into outputs tailored for machine perception—not human viewing—enabling more reliable performance in real-world environments. The market for advanced image sensors is growing rapidly. While the CMOS sensor market is expected to reach $30 billion by 2029, the SPAD market is expanding at a much faster pace, projected to hit $2.55 billion by 2029 with a 20.1% CAGR. This growth reflects increasing demand for high-precision, low-light, and high-speed imaging in autonomous vehicles, robotics, and scientific research. Pi Imaging Technology, a leader in SPAD technology, praised the partnership, with CEO Michel Antolović stating, “By combining our SPAD Alpha’s state-of-the-art hardware with Ubicept’s real-time processing, perception engineers can get the most from what single-photon imaging has to offer.” The Ubicept Toolkit will be available in December 2025. The company is already working with major automakers, AR/VR leaders, and camera manufacturers to push the boundaries of computer vision. With this release, Ubicept aims to make advanced perception technology more accessible, bridging the gap between cutting-edge hardware and the next generation of intelligent systems.

Related Links