National University of Defense Technology

About: National University of Defense Technology is a education organization based out in Changsha, China. It is known for research contribution in the topics: Radar & Synthetic aperture radar. The organization has 39430 authors who have published 40181 publications receiving 358979 citations. The organization is also known as: Guófáng Kēxuéjìshù Dàxué & NUDT.

Temperature sensing using the bandgap-like effect in a selectively liquid-filled photonic crystal fiber.

Abstract: A compact temperature sensor based on a selectively liquid-filled photonic crystal fiber (PCF) is proposed using controlled hole collapse in PCF post-processing. The first ring around the core is filled with liquid of higher refractive index than the matrix, while the outer rings of holes are filled with air. The bandgap (BG)-like effect of the high refractive index ring is analyzed. Absorption loss spectra of the fiber are found to be quite sensitive to the refractive index of liquid when the liquid is lossy. Using the BG-like effect, a fiber temperature sensor is fabricated by selectively injecting a mixture of dimethyl sulfoxide and aqueous gold colloids with a high thermo-optic coefficient to the PCF. Temperature sensitivity up to −5.5 nm/°C is experimentally confirmed.

Real-time mode decomposition for few-mode fiber based on numerical method

Abstract: Today a specific attention has been paid to look into the modal characteristics of the high-power laser beam. And the instantaneous monitoring and analyzing on modal content via the mode decomposition technique will provide a novel route. We implement the first-ever experimental investigation on the real-time mode decomposition technique for few-mode laser beam based on stochastic parallel gradient descent algorithm. It will reduce the cost and the complexity of the mode decomposition system. We have succeeded to decompose the mode spectra as well as calculating the beam quality factor at about 9 Hz monitoring rate, while the high agreement between the measured and reconstructed intensity profiles in each frame indicating the high accuracy and stability during the process. By employing a fiber-squeezing-based polarization controller, the modal content under test can be time-varying automatically.

Rapid Focused Ion Beam Milling Based Fabrication of Plasmonic Nanoparticles and Assemblies via “Sketch and Peel” Strategy

Abstract: Focused ion beam (FIB) milling is a versatile maskless and resistless patterning technique and has been widely used for the fabrication of inverse plasmonic structures such as nanoholes and nanoslits for various applications. However, due to its subtractive milling nature, it is an impractical method to fabricate isolated plasmonic nanoparticles and assemblies which are more commonly adopted in applications. In this work, we propose and demonstrate an approach to reliably and rapidly define plasmonic nanoparticles and their assemblies using FIB milling via a simple “sketch and peel” strategy. Systematic experimental investigations and mechanism studies reveal that the high reliability of this fabrication approach is enabled by a conformally formed sidewall coating due to the ion-milling-induced redeposition. Particularly, we demonstrated that this strategy is also applicable to the state-of-the-art helium ion beam milling technology, with which high-fidelity plasmonic dimers with tiny gaps could be direct...

Thulium/holmium-doped fiber laser passively mode locked by black phosphorus nanoplatelets-based saturable absorber.

Abstract: By coupling black phosphorus (BP) nanoplatelets (NPs) with a fiber-taper evanescent light field, a saturable absorber (SA) based on the BP NPs has been successfully fabricated and used in a thulium/holmium-doped fiber laser as the mode locker. The SA had a modulation depth of ∼9.8% measured at 1.93 μm. A stable mode-locking operation at 1898 nm was achieved with a pulse width of 1.58 ps and a fundamental mode-lock repetition rate of 19.2 MHz. By increasing the pump intensity, phenomena of multi-pulsing operations, including harmonic mode-locked states and soliton bunches, were obtained in the experiment, showing that the BP NPs possess an ultrafast optical response time. This work suggests that the BP NPs-based SA is potentially useful for ultrashort, pulsed laser operations in the eye-safe region of 2 μm.

High-Resolution Encoder–Decoder Networks for Low-Contrast Medical Image Segmentation

Abstract: Automatic image segmentation is an essential step for many medical image analysis applications, include computer-aided radiation therapy, disease diagnosis, and treatment effect evaluation. One of the major challenges for this task is the blurry nature of medical images (e.g., CT, MR, and microscopic images), which can often result in low-contrast and vanishing boundaries. With the recent advances in convolutional neural networks, vast improvements have been made for image segmentation, mainly based on the skip-connection-linked encoder–decoder deep architectures. However, in many applications (with adjacent targets in blurry images), these models often fail to accurately locate complex boundaries and properly segment tiny isolated parts. In this paper, we aim to provide a method for blurry medical image segmentation and argue that skip connections are not enough to help accurately locate indistinct boundaries. Accordingly, we propose a novel high-resolution multi-scale encoder–decoder network (HMEDN), in which multi-scale dense connections are introduced for the encoder–decoder structure to finely exploit comprehensive semantic information. Besides skip connections, extra deeply supervised high-resolution pathways (comprised of densely connected dilated convolutions) are integrated to collect high-resolution semantic information for accurate boundary localization. These pathways are paired with a difficulty-guided cross-entropy loss function and a contour regression task to enhance the quality of boundary detection. The extensive experiments on a pelvic CT image dataset, a multi-modal brain tumor dataset, and a cell segmentation dataset show the effectiveness of our method for 2D/3D semantic segmentation and 2D instance segmentation, respectively. Our experimental results also show that besides increasing the network complexity, raising the resolution of semantic feature maps can largely affect the overall model performance. For different tasks, finding a balance between these two factors can further improve the performance of the corresponding network.