Active mode single-pixel imaging through strong scattering media via least squares conditional generative adversarial networks under low sampling rates

Strong scattering media imaging has become a great challenge in optical imaging field. In this paper, we propose an active mode single-pixel imaging (SPI) method based on a least-squares conditional generation adversarial network that achieves the imaging through strong scattering media (above 100 NTU) at a low sampling rate of 3.52%. The generator of the proposed network uses a U-shaped network structure with an attention mechanism and integrates squeeze-and-excitation blocks and residual blocks, which can learn the target information in the scattering environment better. The least-square loss, content loss, and mean structural similarity loss are used as total loss functions for the first time to stabilize the training process and avoid the gradient disappearance. Simulation and physical experimental results show that the method can effectively improve the image reconstruction quality of SPI under strong scattering conditions at low sampling rates. This method promotes the development of SPI technology and has important applications in optical scattering medium imaging.