Imagia (www.imagia.io), a developer of optical metasurface technology, today announced a breakthrough that allows for optically-accelerated feature detection, enabling image processing operations without power consumption or code. The new technology, called Processing Optics™, allows for the identification of complex features like a human hand or face by means of an array of microscopic optical filters rather than standard, algorithmic on chip image processing. For power-limited and latency-critical applications like edge computing on wearable devices, the ability to extract features of interest quickly and efficiently presents an immense opportunity to enable AI applications at the edge and push the boundaries of device design.
“Processing Optics is a step change in the way we think about extracting information from the world around us,” says Greg Kress, CEO of Imagia. “Searching for objects or patterns within images has traditionally been a slow and computationally intensive task. Now, we get the signal we want at the speed of light, for very little power.”
Imagia’s innovation is rooted in five years of metamaterial optics research from the scientists and engineers at Imagia. After years developing and perfecting metalens technology for optical components used in devices like AR/VR headsets, Imagia has widened its portfolio to explore performing mathematical convolutions directly in optical elements. The technology works by applying a set of mathematical convolutions in an array of optical filters. The light passing through a metalens is steered and transformed by billions of nanoscale components on each Imagia metalens that imparts a hard-coded pattern recognition algorithm to the signal.
Imagia has demonstrated a hand and gesture detector that works with only eight pixels of information and with a response time of only 80 microseconds. By contrast, traditional optics and processing typically take 30-40 milliseconds to process the millions of pixels for digital algorithmic approaches.
By processing the image directly in the optics, Imagia is able to realize a 500x reduction in detection latency for a fraction of the power compared to the traditional method of capturing an image and then processing that data in downstream software. Running at a comparable framerate to a standard image processing system, the Imagia solution consumes less than 1% of the power.
Applications like artificial intelligence and active feature detection in laptops and AR/VR headsets are set to receive outsized benefit from the innovation, which could extend battery life of these devices by 20% or more.
“The optical technology being developed at Imagia allows for compressive sensing of complex features in just a few pixels,” says Bernard Kress, Director of Google AR. “Asking photons to do the electron work allows for fast local processing at lower power in a smaller form factor, key assets of any all-day-use smart glass[es].”
Imagia today is launching an early access program that enables partners to explore solutions and applications running on the Processing Optics platform. The program comes on the heels of the successful traction of Imagia’s closed demos of Processing Optics during 2024 to multinational device makers in markets spanning semiconductors to consumer electronics. Imagia’s first product, a detector module for mobile devices powered by Processing Optics, is expected to launch in 2025.
About Imagia
Imagia is a deep tech startup company that was founded with the vision of applying metamaterial principles to create high-performance optical lenses and filters that are compatible with visible, incoherent light. Imagia optics are fabricated like semiconductors using standard CMOS processes. The technology allows for ultra-fast and ultra-low-power feature detection for next-generation AI devices. Imagia meta-optics enable new possibilities in the design of optical devices and imaging systems that were not previously possible. www.imagia.io
View source version on businesswire.com: https://www.businesswire.com/news/home/20241203535248/en/