Abstract:Conventional active millimeter-wave (MMW) imaging schemes, especially applied in concealed weapon detection, relies on a wide frequency to retrieve information. These methods have resulted in complex and expensive hardware, huge data volumes, and high time consumption when extended to large-scale systems with wide bandwidth. Relying on bandwidth can create difficulties in calibration and alignment. The single-frequency imaging approach is a potential way to solve these existing problems. It can greatly simplify the system and reduce the amount of data and the cost. However, the quality of the reconstructed image will be greatly reduced and not available if only single-frequency echo data are used for imaging in traditional ways. This article proposes an innovative method for single-frequency MMW imaging that can retrieve information and reconstruct high-quality images with only single-frequency echo data. The method is named maximum a posteriori-based single-frequency millimeter imaging method (MAPSM). MAPSM explicitly links the MAP estimation to a physical model by directly incorporating the MAP into the image reconstruction process. It uses data consistency and image prior knowledge as constraints. Furthermore, the MAP-based single-frequency millimeter imaging network (MAPSM-Net) is designed to implement this method. The network consists of an image reconstruction block, a data consistency block, and an image prior knowledge block. We experimentally demonstrate that this proposal can obtain the best-reconstructed image with single-frequency input compared with the conventional reconstruction methods. A comprehensive study of the capabilities is performed to prove that the proposed method obtains the best results in terms of quantitative evaluation and has good generalization ability. This work can be applied to other fields, such as sparse spatial sampling imaging and low-quality MMW image restoration.

Deep Unfolded Approximate Message Passing for Quantitative Acoustic Microscopy Image Reconstruction

A Weighted Hankel Approach and Cramér-Rao Bound Analysis for Quantitative Acoustic Microscopy Imaging

Phase Retrieval from Quantized Measurements Via Approximate Message Passing

A Quantum Denoising-Based Resolution Enhancement Framework for 250-MHz and 500-MHz Quantitative Acoustic Microscopy

Approximate Message Passing in Coded Aperture Snapshot Spectral Imaging

Compressive Imaging via Approximate Message Passing with Image Denoising

BM3D Vector Approximate Message Passing for Radar Coded-aperture Imaging

Compressive Hyperspectral Imaging via Approximate Message Passing

Versatile Denoising-Based Approximate Message Passing for Compressive Sensing.

Reconstructing undersampled photoacoustic microscopy images using deep learning

AMP-Net: Denoising based Deep Unfolding for Compressive Image Sensing

A Model-Based Deep Network for MRI Reconstruction Using Approximate Message Passing Algorithm

Deep image prior for undersampling high-speed photoacoustic microscopy

A Maximum a Posteriori-Based Single-Frequency Millimeter-Wave Image Reconstruction Method.

Accelerated model‐based T1, T2* and proton density mapping using a Bayesian approach with automatic hyperparameter estimation

QAMNet: Fast and Efficient Optical QAM Neural Networks

Tuning-free multi-coil compressed sensing MRI with Parallel Variable Density Approximate Message Passing (P-VDAMP)

Accelerating photoacoustic microscopy by reconstructing undersampled images using diffusion models

UPAMNet: A unified network with deep knowledge priors for photoacoustic microscopy

Switchable deep beamformer for high-quality and real-time passive acoustic mapping

Model-based T1, T2* and Proton Density Mapping Using a Bayesian Approach with Parameter Estimation and Complementary Undersampling Patterns