Abstract:A compact and highly sensitive gas pressure and temperature sensor based on Fabry-Pérot interferometer (FPI) and fiber Bragg grating (FBG) is proposed and demonstrated experimentally in this paper. The theoretical model for pressure and temperature sensing is established. Building on this foundation, a novel micro silicon cavity sensor structure sensitive to pressure is devised downstream of an FBG. The concept of separate measurement and the mechanisms enhancing pressure sensitivity are meticulously analyzed, and the corresponding samples are fabricated. The experimental results indicate that the pressure sensitivity of the sensor is −747.849 nm/MPa in 0—100 kPa and its linearity is 99.7% and it maintains good stability in 150 min. The sensor offers the advantages of compact size, robust construction, easy fabrication, and high sensitivity, making it potentially valuable for micro-pressure application.

Abstract:Reverse design of highly GeO2-doped silica optical fibers with broadband and flat dispersion profiles is proposed using a neural network (NN) combined with a particle swarm optimization (PSO) algorithm. Firstly, the NN model designed to predict optical fiber dispersion is trained with an appropriate choice of hyperparameters, achieving a root mean square error (RMSE) of 9.47×10-7 on the test dataset, with a determination coefficient (R2) of 0.999. Secondly, the NN is combined with the PSO algorithm for the inverse design of dispersion-flattened optical fibers. To expand the search space and avoid particles becoming trapped in local optimal solutions, the PSO algorithm incorporates adaptive inertia weight updating and a simulated annealing algorithm. Finally, by using a suitable fitness function, the designed fibers exhibit flat group velocity dispersion (GVD) profiles at 1 400—2 400 nm, where the GVD fluctuations and minimum absolute GVD values are below 18 ps∙nm-1∙km-1 and 7 ps∙nm-1∙km-1, respectively.

Abstract:To fulfill the need for acquiring three-dimensional (3D) objects with more realistic textures and depth information, this study proposes a method based on near-infrared laser, combined with dual camera field of view center correction and binocular stereo calibration, to precisely capture the target surface texture. Furthermore, we constructed a verification system using standard industrial cameras and line lasers, achieving the generation of binocular line laser point cloud real textures. Experiments conducted within a 400 mm to 600 mm testing range achieved a reconstruction accuracy of 0.047 2 mm and reduced the texture mapping error to 0.323 4 pixel, proving the effectiveness of this method and providing a high-precision, low-cost solution for 3D point cloud model texture mapping.

Abstract:Sign language dataset is essential in sign language recognition and translation (SLRT). Current public sign language datasets are small and lack diversity, which does not meet the practical application requirements for SLRT. However, making a large-scale and diverse sign language dataset is difficult as sign language data on the Internet is scarce. In making a large-scale and diverse sign language dataset, some sign language data qualities are not up to standard. This paper proposes a two information streams transformer (TIST) model to judge whether the quality of sign language data is qualified. To verify that TIST effectively improves sign language recognition (SLR), we make two datasets, the screened dataset and the unscreened dataset. In this experiment, this paper uses visual alignment constraint (VAC) as the baseline model. The experimental results show that the screened dataset can achieve better word error rate (WER) than the unscreened dataset.

Abstract:The unmanned aerial vehicle (UAV) images captured under low-light conditions are often suffering from noise and uneven illumination. To address these issues, we propose a low-light image enhancement algorithm for UAV images, which is inspired by the Retinex theory and guided by a light weighted map. Firstly, we propose a new network for reflectance component processing to suppress the noise in images. Secondly, we construct an illumination enhancement module that uses a light weighted map to guide the enhancement process. Finally, the processed reflectance and illumination components are recombined to obtain the enhancement results. Experimental results show that our method can suppress the noise in images while enhancing image brightness, and prevent over enhancement in bright regions. Code and data are available at https://gitee.com/ baixiaotong2/uav-images.git.

Abstract:We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hierarchical efficient multi-scale attention (H-EMA) module is designed for lightweight feature extraction, achieving outstanding performance at a relatively low cost. Secondly, an improved EfficientNetV2 block is used to integrate information from different scales better and enhance inter-layer message passing. Furthermore, introducing the convolutional block attention module (CBAM) enhances the model’s perception of critical features, optimizing its generalization ability. Lastly, Focal Loss is introduced to adjust the weights of complex samples to address the issue of imbalanced categories in the dataset, further improving the model’s performance. The model achieved 96.11% accuracy on the intertidal marine organism dataset of Nanji Islands and 84.78% accuracy on the CIFAR-100 dataset, demonstrating its strong generalization ability to meet the demands of oceanic biological image classification.

Abstract:To improve image quality under low illumination conditions, a novel low-light image enhancement method is proposed in this paper based on multi-illumination estimation and multi-scale fusion (MIMS). Firstly, the illumination is processed by contrast-limited adaptive histogram equalization (CLAHE), adaptive complementary gamma function (ACG), and adaptive detail preserving S-curve (ADPS), respectively, to obtain three components. Then, the fusion-relevant features, exposure, and color contrast are selected as the weight maps. Subsequently, these components and weight maps are fused through multi-scale to generate enhanced illumination. Finally, the enhanced images are obtained by multiplying the enhanced illumination and reflectance. Compared with existing approaches, this proposed method achieves an average increase of 0.81% and 2.89% in the ​structural similarity index measurement (SSIM) and peak signal-to-noise ratio (PSNR), and a decrease of 6.17% and 32.61% in the natural image quality evaluator (NIQE) and gradient magnitude similarity deviation (GMSD), respectively.

Abstract:Considering the three-dimensional (3D) U-Net lacks sufficient local feature extraction for image features and lacks attention to the fusion of high- and low-level features, we propose a new model called 3DMAU-Net based on the 3D U-Net architecture for liver region segmentation. Our model replaces the last two layers of the 3D U-Net with a sliding window-based multilayer perceptron (SMLP), enabling better extraction of local image features. We also design a high- and low-level feature fusion dilated convolution block that focuses on local features and better supplements the surrounding information of the target region. This block is embedded in the entire encoding process, ensuring that the overall network is not simply downsampling. Before each feature extraction, the input features are processed by the dilated convolution block. We validate our experiments on the liver tumor segmentation challenge 2017 (Lits2017) dataset, and our model achieves a Dice coefficient of 0.95, which is an improvement of 0.015 compared to the 3D U-Net model. Furthermore, we compare our results with other segmentation methods, and our model consistently outperforms them.

Abstract:Although ray tracing produces high-fidelity, realistic images, it is considered computationally burdensome when implemented on a high rendering rate system. Perception-driven rendering techniques generate images with minimal noise and distortion that are generally acceptable to the human visual system, thereby reducing rendering costs. In this paper, we introduce a perception-entropy-driven temporal reusing method to accelerate real-time ray tracing. We first build a just noticeable difference (JND) model to represent the uncertainty of ray samples and image space masking effects. Then, we expand the shading gradient through gradient max-pooling and gradient filtering to enlarge the visual receipt field. Finally, we dynamically optimize reusable time segments to improve the accuracy of temporal reusing. Compared with Monte Carlo ray tracing, our algorithm enhances frames per second (fps) by 1.93× to 2.96× at 8 to 16 samples per pixel, significantly accelerating the Monte Carlo ray tracing process while maintaining visual quality.