Under-sampled retina-inspired Hadamard ghost imaging enhanced by cake-cutting ordering

Inspired by the foveated imaging of the human retina, retina-inspired variable-resolution speckle patterns have been introduced into ghost imaging to efficiently capture regions of interest. We propose under-sampled retina-inspired Hadamard ghost imaging enhanced by cake-cutting ordering. By arranging the projected speckles in ascending order based on two-dimensional connected components, the imaging results transition from coarse to fine as the sampling rate increases, aligning with human visual characteristics. Further optimization of low-contrast imaging is achieved through lateral inhibition enhancement. Simulations and experiments demonstrate that the optimized speckle patterns integrate the information compression of biomimetic structures, the rapid generation and inverse transformation of Hadamard speckles, and the high efficiency of cake-cutting ordering. This enables high-resolution imaging of targets within the field of view at a low sampling rate and computational cost. The method complements existing retina-inspired ghost imaging techniques and holds promise for advancing the development of ghost imaging.