New Generation of SPAD-Based Image Sensors

Digital imaging is changing, but most SPAD sensors only count photons. What if they could process data at the pixel level in real time?

Singular Photonics, a spinout from the University of Edinburgh, is among the first to integrate advanced computation directly into SPAD-based image sensing. The company enables in-pixel and cross-pixel storage and processing at extremely low light levels, revealing details of photon interactions that were previously undetectable.

SPAD sensors use the “avalanche” effect in semiconductors to convert light into electrical signals without needing cooling or amplification. While most commercial SPAD sensors focus on time-resolved photon counting, Singular’s approach integrates computational layers beneath 3D-stacked SPAD sensors. This is similar to how FPGAs and GPUs revolutionized parallel computing with high-speed, localized processing.

– Advertisement –

SPAD sensors are the future of digital imaging, but their commercial use has mostly been limited to counting photons over time. Singular is changing that by embedding digital computation directly at the pixel level—where the photons arrive.

By capturing both depth and temporal data to create 4D images, Singular’s sensors extract more information from light. This supports applications in consumer electronics, automotive technology, scientific research, and medical imaging. The company’s SPAD sensors function as 3D-stacked computational engines, capable of real-time photon counting, timing, and in-pixel statistical analysis.

The company is launching two commercial sensors:

• Andarta, developed with Meta, is a compact, high-sensitivity sensor designed for medical imaging. It supports multiple operation modes, including in-pixel autocorrelation, and brings SPAD technology closer to wearables. One key application is monitoring cerebral blood flow by detecting subtle light fluctuations in tissue—at depths beyond the reach of existing sensors.
• Sirona, a 512-pixel SPAD-based line sensor, is designed for time-correlated single-photon counting (TCSPC). It enables applications such as Raman spectroscopy, fluorescence lifetime imaging microscopy (FLIM), time-of-flight imaging, and quantum computing. With on-chip histogramming and time binning, Sirona has the potential to advance spectroscopy technologies significantly.

For more information, click here.