Metalens array for complex-valued optical discrete Fourier transform
Photonic computing has emerged as a promising platform for accelerating computational tasks with high degrees of parallelism, such as image processing and neural network. We present meta-DFT (discrete Fourier transform), a single layer metasurface device, designed to perform optical complex-to-complex DFT with O(N) time complexity. One critical challenge in free-space analog optical computing is to control the measurement error. Our scheme addresses this issue by focusing light on spatially separated focal points and reconstructing the complex phase, which enable error correction. We systematically evaluate the device's performance using input vectors with random complex amplitudes and phases, to demonstrate its robust accuracy. Our findings pave the way towards advancement of metasurface-based computation, offering a robust framework that is readily extensible to an arbitrary complex-valued matrix-vector multiplication (MVM).