In this article, we demonstrate an efficient terahertz single-pixel imaging system incorporating deep learning networks. Experimental results show that by combining a Hadamard single-pixel imaging system with the deep learning network, the sampling time per pattern can be reduced to 1/20 of the conventional system and the number of Hadamard patterns can be reduced to 10% of the pixels while maintaining high image quality with acceptable signal-to-noise ratio above 20 dB and structural similarity of more than 0.85. We thus expect this article to advance the development of a real-time terahertz single-pixel imaging system and promote its applications.

We demonstrate terahertz single-pixel imaging is improved by using a photomodulator based on silicon passivated with SiO ${}_2$ . By exploring various SiO ${}_2$ thicknesses, we show that the modulation factor of the as-fabricated terahertz photomodulator can reach 0.9, three times that based on bare silicon. This improvement originates from chemical passivation, as well as anti-reflection. Single-pixel imaging experiments based on the compressed sensing method show that reconstructed images adopting the new photomodulator have better quality than the conventional terahertz modulator based on bare silicon. Since the passivation process is routine and low cost, we expect this work will reduce the cost of terahertz photomodulator and single-pixel THz imaging, and advance their applications.

We propose and demonstrate a single-pixel imaging method based on deep learning network enhanced singular value decomposition. The theoretical framework and the experimental implementation are elaborated and compared with the conventional methods based on Hadamard patterns or deep convolutional autoencoder network. Simulation and experimental results show that the proposed approach is capable of reconstructing images with better quality especially under a low sampling ratio down to 3.12%, or with fewer measurements or shorter acquisition time if the image quality is given. We further demonstrate that it has better anti-noise performance by introducing noises in the SPI systems, and we show that it has better generalizability by applying the systems to targets outside the training dataset. We expect that the developed method will find potential applications based on single-pixel imaging beyond the visible regime.

We numerically investigate the thermal effects in a cornea illuminated by terahertz radiation. By modifying the bioheat and Arrhenius equations, we studied the heat-transfer and temperature distributions in the corneal tissue, and evaluated the potential thermal damage. The influence of the beam radius and power density are discussed. We also estimated the effective cornea-collagen shrinkage region, and evaluated the degree of thermal damage in the cornea. We expect this work to open up a novel effective and safe thermal-treatment approach based on THz radiation for cornea reshaping in the field of ophthalmology.

Terahertz single-pixel imaging technology, which utilizes the unique properties of terahertz waves in combination with a single-pixel detector, offers an effective solution to the scarcity of high-performance focal-plane array detectors in the terahertz regime. However, its widespread application has been hindered primarily by low imaging frame rates. In this study, we propose a high-speed terahertz single-pixel imaging approach based on a vertical-cavity surface-emitting laser array. By integrating a custom-designed 32×32 infrared vertical-cavity surface-emitting laser module with field-programmable gate array-based control and optimized Hadamard encoding, our system overcomes the mechanical switching limitations of conventional digital micromirror devices. Experimental results show the system achieves real-time imaging at 25 frames per second under full sampling. Furthermore, under undersampling conditions, the frame rate reaches 50 fps. Compared to existing methods, our approach improves the imaging speed by an order of magnitude at equivalent resolution while preserving full sampling capability. This advancement offers a fast and efficient solution for dynamic terahertz imaging, with strong potential for applications in security screening, biomedical diagnostics, and beyond.

Terahertz (THz) single-pixel imaging has emerged as a significant research direction due to its cost-effectiveness, low system complexity, and high scalability. However, practical implementation is hindered by a critical trade-off between imaging speed and quality, stemming from the combined limitations of modulator response speeds and reconstruction algorithms. Addressing this challenge, we propose a physics-enhanced deep learning framework for video-rate THz single-pixel imaging. By utilizing a high-speed vertical-cavity surface-emitting laser array to pump a high-resistivity silicon wafer, we achieve high-speed THz modulation and remove the refresh-rate limitation associated with commonly used DMD-based optical pattern generation. To bridge the gap between hardware addressing constraints and the requirements for complex encoding masks, a low-rank matrix synthesis strategy is integrated into the neural network training to jointly optimize mask distribution and image reconstruction. Both simulation and experimental results demonstrate that this synergistic mechanism outperforms pure data-driven algorithms and traditional basis scanning methods, enabling real-time THz video capture at 50 f/s with a resolution of 32 × 32 pixels. Even at a 25% sampling ratio, the system achieves high-fidelity reconstruction, effectively balancing speed and quality. This hardware–software co-designed approach paves the way for real-time THz applications in industrial and scientific research while establishing a new paradigm for high-speed dynamic imaging across other spectral regimes.