Real-time generation of fully optimized holograms for optical trapping applications

Paper in proceeding, 2011

We have implemented several algorithms for hologram generation, aimed for holographic optical tweezers applications, using the parallel computing architecture CUDA. We compare required computation time for different implementations of the Gerchberg-Saxton algorithm and provide guidelines for choosing the best suited version with respect to the application. We also show that additional calculations, compensating for limitations in the used spatial light modulator and optical system, can be included in the hologram generating software with little or no loss in computational speed.