Efficient Kronecker-based deep unfolding network for single-pixel imaging with hardware implementation

TY - JOUR

T1 - Efficient Kronecker-based deep unfolding network for single-pixel imaging with hardware implementation

AU - Zhou, Chang

AU - Cao, Jie

AU - Yao, Haifeng

AU - Hao, Qun

AU - Zhang, Haoyu

AU - Deng, Loulou

N1 - Publisher Copyright: © 2026 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement.

PY - 2026/3/9

Y1 - 2026/3/9

N2 - Single-pixel imaging (SPI) is a novel computational imaging technique. Recently, deep learning has achieved remarkable improvements in SPI reconstruction in both quality and speed. However, the rapidly increasing computational costs at large image sizes remain a key obstacle to practical SPI. In this work, we propose the K-CPP Net, a gradient-descent-based deep unfolding network that couples a Kronecker-based SPI model with a Chambolle-Pock-inspired primal-dual iterative optimization algorithm for efficient SPI reconstruction. Specifically, we avoid explicit vectorization of the large sensing matrix by representing the matrix as the Kronecker product of two learnable subsampling matrices, and design a lightweight denoising module based on residual convolutional blocks and efficient channel-wise self-attention. This method substantially reduces the computational cost, enabling full-image SPI training and inference while effectively suppressing distortion and blur and preserving fine structural details. Moreover, we implement K-CPP Net on an FPGA via an optimized accelerator architecture, providing a reference design for real-world SPI applications. Experimental results demonstrate that the proposed method achieves superior reconstruction quality while offering markedly improved computational efficiency.

AB - Single-pixel imaging (SPI) is a novel computational imaging technique. Recently, deep learning has achieved remarkable improvements in SPI reconstruction in both quality and speed. However, the rapidly increasing computational costs at large image sizes remain a key obstacle to practical SPI. In this work, we propose the K-CPP Net, a gradient-descent-based deep unfolding network that couples a Kronecker-based SPI model with a Chambolle-Pock-inspired primal-dual iterative optimization algorithm for efficient SPI reconstruction. Specifically, we avoid explicit vectorization of the large sensing matrix by representing the matrix as the Kronecker product of two learnable subsampling matrices, and design a lightweight denoising module based on residual convolutional blocks and efficient channel-wise self-attention. This method substantially reduces the computational cost, enabling full-image SPI training and inference while effectively suppressing distortion and blur and preserving fine structural details. Moreover, we implement K-CPP Net on an FPGA via an optimized accelerator architecture, providing a reference design for real-world SPI applications. Experimental results demonstrate that the proposed method achieves superior reconstruction quality while offering markedly improved computational efficiency.

UR - https://www.scopus.com/pages/publications/105032362847

U2 - 10.1364/OE.586333

DO - 10.1364/OE.586333

M3 - Article

AN - SCOPUS:105032362847

SN - 1094-4087

VL - 34

SP - 8475

EP - 8491

JO - Optics Express

JF - Optics Express

IS - 5

ER -