Encoding and decoding of multidimensional optical field modulation in holographic data storage

Holographic data storage (HDS) offers a promising route toward high-capacity and high-throughput optical information storage. However, its achievable storage density is often limited by the dimensionality of the encoded optical field. Light inherently possesses multiple physical degrees of freedom, including amplitude, phase, and polarization, which provide an opportunity to substantially increase the information capacity of a single data page if they can be efficiently and reliably exploited. In this work, we investigate multidimensional information encoding and decoding in the HDS by fully utilizing these three fundamental properties of light. For encoding, a full three-dimensional modulation scheme is established by controlling the complex amplitude distributions of two orthogonal polarization states and employing a double-phase hologram strategy, which allows a single phase-only spatial light modulator to jointly encode amplitude, phase, and polarization information within the optical field. For decoding, a convolutional neural network, named TriDecode-Net, is designed to simultaneously retrieve three-dimensional information directly from diffraction intensity images. This approach enables synchronous recording and retrieval of multidimensional optical information, significantly increasing the information content of a single holographic data page compared with conventional holographic encoding approaches. Beyond data storage, the proposed multidimensional modulation and reconstruction framework also provides opportunities for optical encryption, anti-counterfeiting, high-capacity optical communication, and the generation and control of structured light fields.