Hua Bai

Wind turbine blade defect detection requires accurate identification of small and irregular defects while maintaining low computational cost for practical inspection scenarios. However, lightweight detectors often suffer from insufficient local feature extraction, limited multiscale feature fusion, and weak responses to critical defect regions. To address these issues, this study proposes a Receptive-Field-Enhanced You Only Look Once model (RFE-YOLO), a lightweight defect detection model based on You Only Look Once version 10 nano (YOLOv10n).The proposed model introduces three task-oriented improvements. First, C2f-RFAConv is embedded into the backbone to enhance receptive field aware local feature representation for fine grained defects. Second, a Compact Cross-scale Feature Fusion Module, termed CCFM, is designed in the neck to improve the integration of low-level detail information and high-level semantic features with reduced computational complexity. Third, an Efficient Local Attention module is inserted before the detection head to strengthen defect-related spatial responses after feature fusion. Experiments were conducted on a wind turbine blade defect dataset containing three categories, namely Crack, Oil leakage, and Peel. The results show that RFE-YOLO achieves 89.9% mean Average Precision at an Intersection over Union threshold of 0.5, namely mAP@0.5, and 64.73% mAP@0.5:0.95. Compared with YOLOv10n, RFE-YOLO improves mAP@0.5 by 2.8 percentage points while reducing the number of parameters from 2.70M to 1.91M and giga floating point operations from 8.4 to 5.3. The inference speed reaches 88.8 frames per second on an NVIDIA GeForce RTX 3090 GPU. These results indicate that RFE-YOLO achieves a favorable balance between detection accuracy and model efficiency under the current experimental setting.

Honey products adulterated with syrup need to be availably distinguished from the true ones and quantified by modern method, for instance three-dimensional fluorescence spectroscopy. The spectra from eight types of honey showed that a linear relationship existed between fluorescence intensity and honey solution concentration in two excitation wavelength ranges of 240–320 nm and 320–360 nm when honey was diluted to 1 % ( V : V ) or less. Based on the linearity, a way was proposed to estimate the purity of adulterated honey by employing the least squares model to compare the spectra of adulterated honey with honey and syrup on a computer, after they were properly diluted. Honey artificially mixed with syrup in a ratio of 3:7 ( V : V ) was evaluate to be 31.6 % and the error was only 1.6 %. The method of quantitation on adulterated honey has the characteristics of convenience and accuracy.

Accurate segmentation of medical images is essential for clinical decision-making, and deep learning techniques have shown remarkable results in this area. However, existing segmentation models that combine transformer and convolutional neural networks often use skip connections in U-shaped networks, which may limit their ability to capture contextual information in medical images. To address this limitation, we propose a coordinated mobile and residual transformer UNet (MRC-TransUNet) that combines the strengths of transformer and UNet architectures. Our approach uses a lightweight MR-ViT to address the semantic gap and a reciprocal attention module to compensate for the potential loss of details. To better explore long-range contextual information, we use skip connections only in the first layer and add MR-ViT and RPA modules in the subsequent downsampling layers. In our study, we evaluated the effectiveness of our proposed method on three different medical image segmentation datasets, namely, breast, brain, and lung. Our proposed method outperformed state-of-the-art methods in terms of various evaluation metrics, including the Dice coefficient and Hausdorff distance. These results demonstrate that our proposed method can significantly improve the accuracy of medical image segmentation and has the potential for clinical applications. Illustration of the proposed MRC-TransUNet. For the input medical images, we first subject them to an intrinsic downsampling operation and then replace the original jump connection structure using MR-ViT. The output feature representations at different scales are fused by the RPA module. Finally, an upsampling operation is performed to fuse the features to restore them to the same resolution as the input image.

This Special Issue entitled “Optoelectronic Materials, Devices, and Applications” is devoted to gathering a broad array of research papers on the latest advances in the development of optoelectronic materials and devices of semiconductors, fiber optics, power electronics, microwaves, and terahertz

Acoustic neuroma is one of the most common tumors in the cerebellopontine angle area. Patients with acoustic neuroma have clinical manifestations of the cerebellopontine angle occupying syndrome, such as tinnitus, hearing impairment and even hearing loss. Acoustic neuromas often grow in the internal auditory canal. Neurosurgeons need to observe the lesion contour with the help of MRI images, which not only takes a lot of time, but also is easily affected by subjective factors. Therefore, the automatic and accurate segmentation of acoustic neuroma in cerebellopontine angle on MRI is of great significance for surgical treatment and expected rehabilitation. In this paper, an automatic segmentation method based on Transformer is proposed, using TransUNet as the core model. As some acoustic neuromas are irregular in shape and grow into the internal auditory canal, larger receptive fields are thus needed to synthesize the features. Therefore, we added Atrous Spatial Pyramid Pooling to CNN, which can obtain a larger receptive field without losing too much resolution. Since acoustic neuromas often occur in the cerebellopontine angle area with relatively fixed position, we combined channel attention with pixel attention in the up-sampling stage so as to make our model automatically learn different weights by adding the attention mechanism. In addition, we collected 300 MRI sequence nuclear resonance images of patients with acoustic neuromas in Tianjin Huanhu hospital for training and verification. The ablation experimental results show that the proposed method is reasonable and effective. The comparative experimental results show that the Dice and Hausdorff 95 metrics of the proposed method reach 95.74% and 1.9476 mm respectively, indicating that it is not only superior to the classical models such as UNet, PANet, PSPNet, UNet++, and DeepLabv3, but also show better performance than the newly-proposed SOTA (state-of-the-art) models such as CCNet, MANet, BiseNetv2, Swin-Unet, MedT, TransUNet, and UCTransNet.

Photodynamic therapy (PDT) is considered a promising noninvasive therapeutic strategy in biomedicine, especially by utilizing low-level laser therapy (LLLT) in visible and near-infrared spectra to trigger biological responses. The major challenge of PDT in applications is the complicated and time-consuming biological methodological measurements in identification of light formulas for different diseases. Here, we demonstrate a rapid and label-free identification method based on artificial intelligence (AI)-assisted terahertz imaging for efficient light formulas in LLLT of acute lung injury (ALI). The gray histogram of terahertz images is developed as the biophysical characteristics to identify the therapeutic effect. Label-free terahertz imaging is sequentially performed using rapid super-resolution imaging reconstruction and automatic identification algorithm based on a voting classifier. The results indicate that the therapeutic effect of LLLT with different light wavelengths and irradiation times for ALI can be identified using this method with a high accuracy of 91.22% in 33 s, which is more than 400 times faster than the biological methodology and more than 200 times faster than the scanning terahertz imaging technology. It may serve as a new tool for the development and application of PDT.

Terahertz (THz) photonic crystal (PC) waveguides show promise as an efficient and versatile waveguiding platform for communication, sensing, and imaging. However, low-frequency THz PC waveguides with a low-cost and easy fabrication remain challenging. To address this issue, a THz PC waveguide with a lilac-shaped defect has been designed and fabricated by 3D printing based on stereolithography (SLA). The reflection and transmission characteristics of the proposed waveguide have been analyzed using the finite difference frequency domain (FDFD) method. The waveguide spectral response is further optimized by changing the distance of the lilac-shaped resonant cavities. Consistent with the results of numerical modeling, the measured results show that the waveguide performs a resonant reflection in the region of 0.2 to 0.3 THz and low-pass transmission in the 6G mobile communication window. Furthermore, in order to characterize the performance of the proposed waveguide, parameters have been analyzed, including the Q factor, resonant frequency, and bandwidth. This work supplies a novel pathway for the design and fabrication of a low-frequency THz PC waveguide with potential applications in communication, sensing, and imaging.

Three-dimensional compound measurement within a volume of interest is of great importance in industrial manufacturing and the biomedical field. However, there is no current method that can simultaneously perform spatial localization and 3D measurement in a non-scanning manner as well as the identification of material in a volume. In this paper, an Autostereoscopic-Raman Spectrometry-based (ARS) three-dimensional measurement system is proposed. The target object in a large depth range is initially positioned by the autostereoscopic 3D measurement method, and then the accurate position information is cross-checked and obtained by combining the spectral signal. Meanwhile, the spectral signal at the precise excitation position guided by the autostereoscopic signal also carries the material composition information. In order to verify the proposed ARS method, an associated measurement system was developed, and experimental studies of detecting various fibers of different depths in multi-layer glass structure were conducted. The spatial locations and dimensional information of multiple different targets can be measured in a volume, and their material can also be identified at the same time. The average error between the calculated position processed by the ARS system and the actual spatial position is within sub-micron levels, and the success rate of spectrum acquisition reaches 98%.

Hollow-core terahertz (THz) fibers have attracted a lot of research interest due to the low loss and easy inner modification with functional materials. Liquid water has unique properties in the THz region and has been widely investigated in THz emission, sensing, and devices. In this paper, a hollow-core THz Bragg fiber with a water defect layer is proposed. The finite element method is used to verify and analyze the tunable temperature characteristic of the water-filled THz fiber. The numerical analysis results show that the confinement loss and the low-frequency side of the dip near 0.5 THz can be controlled by the temperature of the liquid water. The temperature sensitivity of the THz fiber is obtained at 0.09614 dB·m⁻¹/K at 0.45 THz with a high core power fraction up to 98%. The proposed THz fiber has potential applications in THz interaction with liquid and THz tunable devices.