US’ light-powered AI camera processes images 99.4% faster without electricity
SMRTR summary
A groundbreaking AI camera developed by researchers at Princeton and Washington universities uses light to process images 99.4% faster than traditional methods. The compact device incorporates metasurface lenses functioning as an optical neural network, enabling ultra-fast, low-energy image recognition. This approach performs hundreds of millions of calculations instantly, using less than 1% of the computing power required by conventional systems. The technology has potential applications beyond image processing and could revolutionize computer vision and machine intelligence.
SMRTR provides this summary for quick context. The original article belongs to Interesting Engineering.
Read the original article