3D information transmission of a computer-generated hologram using a quantum compensation hybrid neural network

Abstract

A computer-generated hologram compression and transmission system is designed. The system uses a hybrid neural network based on quantum compensation to compress and decompress computer-generated holograms and uses a quantization method and Huffman coding to encode the compressed holograms for transmission. The receiving side decodes and decompresses the data stream to obtain the recovered hologram, which can be used to reconstruct the original object information for 3D image display. Our numerical experiments show that quantum compensation can improve the quality of the reconstructed image. Furthermore, we introduce a quantum compensation factor to describe the strength of quantum compensation and verify experimentally that the quantum compensation module may drive the conventional neural network to converge in fewer training epochs.