Binary wavefront optimization using a simulated annealing algorithm

Abstract

We propose an idea using a simulated annealing algorithm for amplitude modulation to focus light through disordered media. Using 4096 independently controlled segments of an incident wavefront, the intensity of the target signal is enhanced 73 times over the original intensity of the same output channel. The simulated annealing algorithm and existing amplitude control algorithms for focusing through scattering media are compared experimentally. It is found that the simulated annealing algorithm achieves the highest enhancement when the number of iterations required for optimization is the same.

Wavefront shaping: A versatile tool to conquer multiple scattering in multidisciplinary fields

“… 55 , 78 , 79 The input-output mapping is inevitably relaxed for adaptive modulation, and feedback-based optimization is therefore required. These optimization algorithms, including evolutional algorithms, 79 , 80 , 81 , 82 , 83 , 84 , 85 artificial intelligence algorithms, 86 , 87 , 88 and their combination, 89 , 90 can instantly modify the modulation patterns while screening the feedback variations. Encountering stronger variations (e.g., a dynamic medium), optimization with physics prior 91 , 92 , 93 works more efficiently, which quantifies the error in the optimized wavefront for optical focusing, while previously the number of to-be-corrected pixels on a spatial light modulator (SLM) is empirically guessed.…”

Optimal efficiency of focusing diffused light through scattering media with iterative wavefront shaping

“…In this technique, the wavefront of incident light is iteratively modulated by a spatial light modulator (SLM) according to the feedback signals from a detector (e.g., camera or photodetector) placed behind the scattering medium or a guidestar (e.g., fluorescent or photoacoustic emission as a virtual guidestar 40,41 ) within the scattering medium; hence, the optical distortions can be compensated and an optical focus or the desired output field can be obtained at the target location. Several metaheuristic optimization algorithms, such as genetic algorithm (GA), [17][18][19][20][21] particle swarm optimization (PSO), [22][23][24][25][26] simulated annealing algorithm (SA), 27,28 bat algorithm (BA), 31 ant colony optimization (ACO), 32 and separable natural evolution strategies (SNES), 33 have been demonstrated for successful optical focusing inside/through scattering media. Aiming for faster and better optimization results, researchers have improved different algorithms primarily in three ways: optimizing parameters, introducing variants of the algorithms, such as microgenetic algorithm (μGA), 19 particle swarm optimization with mutation, 26 and improved ant colony optimization (IACO); 32 and exploiting hybrid algorithms, such as parameter-free algorithm (PFA), a combination of genetic algorithm and bat algorithm, 30 and Genetic Neural Network (GeneNN), a hybrid of GA and deep neural networks.…”

Approaching the optimum focusing of diffused light through scattering media with iterative wavefront shaping by a probability-based algorithm

“…However, due to the inhomogeneities of refractive index in biological tissue, photons are multiply scattered in tissue sample and the amount of ballistic photons decays exponentially with increasing propagation depth, limiting high-resolution optical imaging to a depth of ~1 mm beneath the skin or tissue surface 1 . Thanks to the emergence of optical wavefront shaping techniques, scattering-induced wavefront distortions nowadays can be compensated via various approaches, such as optical phase conjugation [2][3][4][5][6][7][8] , transmission matrix method 9-12 , iterative optimization [13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28][29] , and artificialintelligence-assisted methods [30][31][32][33] , allowing optical focusing or imaging inside/through scattering media. Iterative optimization approaches are widely adopted because they are straightforward and less technically demanding.…”

“…In this technique, the wavefront of incident light is iteratively modulated by a spatial light modulator (SLM) according to the feedback signals from a detector (e.g., camera or photodetector) placed behind the scattering medium or a guidestar (e.g., fluorescent or photoacoustic emission as a virtual guidestar 34,35 ) within the scattering medium, so that the optical distortions can be compensated and an optical focus or a desired output field can be obtained at the target location. Several metaheuristic optimization algorithms, such as genetic algorithm (GA) [15][16][17][18] , particle swarm optimization (PSO) [19][20][21][22][23] , simulated annealing algorithm (SA) 24,25 , bat algorithm (BA) 28 , and separable natural evolution strategies (SNES) 29 , have been demonstrated for successful optical focusing inside/through scattering media. Aiming for faster and better optimization results, researchers have improved different algorithms primarily in three ways: optimizing parameters, introducing variants of the algorithms (e.g., microgenetic algorithm (μGA) 17 and particle swarm optimization with mutation 23 ), and exploiting hybrid algorithms (e.g., parameter-free algorithm (PFA), a combination of genetic algorithm and bat algorithm 27 , and Genetic Neural Network (GeneNN), a hybrid of GA and deep neural networks 31 ).…”