Explore the vast range of titles available.

The detection of mirrors and glass, which possess unique optical surface properties, has garnered significant attention in recent years. Due to their reflective and transparent nature, these surfaces are often difficult to distinguish from their surrounding environments, posing substantial challenges even for advanced deep learning models tasked with performing such detection. Current research primarily relies on complex network models that learn and fuse different modalities of images, such as RGB, depth, and thermal, to achieve mirror and glass detection. However, these approaches often overlook the inherent limitations in the raw data caused by sensor deficiencies when facing mirrors and glass surfaces. To address this issue, we applied mathematical and physical methods, such as three-point plane determination and steady-state heat conduction in two-dimensional planes, along with an RGB enhancement module, to reconstruct RGB, depth, and thermal data for mirrors and glass in two publicly available datasets: an RGB-D mirror detection dataset and an RGB-T glass detection dataset. Additionally, we synthesized four enhanced and ideal datasets. Furthermore, we propose a double weight Mamba fusion network (DWMFNet) that strengthens the model’s global perception of image information by extracting low-level clue weights and high-level contextual weights from the input data using the prior fusion feature extraction module (PFFE) and the deep fusion feature guidance module (DFFG). This is complemented by the Mamba module, which efficiently captures long-range dependencies, facilitating information complementarity between multi-modal features. Extensive experiments demonstrate that our data enhancement method significantly improves the model’s capability in detecting mirrors and glass surfaces.

In order to solve these problems, such as forgetting to take medicine, wrong dosage or type of medicine, an intelligent medicine box is designed. The intelligent medicine box is composed of a piston dispensing mechanism, a platform moving mechanism and a main body mechanism, etc. Close cooperation between various parts, realize the functions of storing medicines, automatically dispensing medicines, intelligent reminders, intelligent detection and intelligent sterilization. Through the simulation in solidworks, the mechanical performance simulation analysis. The results show that the deformation of the crank is very small during the whole process of dispensing medicines, there is no concentrated stress, and it has no effect on the dispensing efficiency during the dispensing process. The designed intelligent medicine box is not only for the elderly, but also suitable for all kinds of people. It can promptly remind patients to take drugs, reduce the occurrence of problems such as wrong medication and forgetting to take drugs. It is user-friendly in design, simple and convenient to operate, helps smart medical treatment, eases the pressure of medical care, and has good market application value.

Nowadays, Automatic metal surface defect recognition is an important research direction in the field of surface defect recognition, and more convolution neural network algorithms are applied in this field. However, with the deepening of network layers, network degradation will occur. We propose a ResNet method for classifying metal surface defects. After experimental testing, we use ResNet34 to build an identification network. After training 300 epoch of the network using the NEU surface defect data set, the convergence of the network is very good. The accuracy of test set verification is 93.67% and higher than that of other surface defect recognition algorithms. Also, we can deepen the number of layers ResNet the network without worrying about network degradation.

How to optimize the quality of the truss so that the truss meets the load-bearing requirements has always been a key issue of research. In this paper, a genetic algorithm (GA) and finite element analysis (FEA) based optimization method is proposed for the size and topology of a space truss. According to the results of space truss, four kinds of components are divided and their cross-sectional areas are optimized respectively. The coded value on the chromosome in GA is used to represent the truss topology, and the adjacent value represents the adjacent truss structure. The nodal displacement and member stress of the truss are solved by finite element method to ensure the safety of the truss. All work is done using python.

Truss structure is widely used in aircraft design. In this paper, the lightweight design of a certain type of aircraft truss structure is carried out through topology analysis and size optimization to improve the performance of the aircraft. The original truss model is established based on ABAQUS, and its static strength is checked and analyzed. According to the design domain of the original model, a topology optimization method is used to obtain a new material distribution. Then the section size of each truss member is taken as the optimization variable, the maximum deformation of the structure is taken as the constraint, and the overall model volume is the minimum as the optimization goal, the genetic algorithm is used to obtain the new structure size. The results show that the truss members after the second optimization can meet the requirements of use and have a 75.15% reduction in mass compared to the original structure, which verifies that the method is feasible and provides a new idea for the lightweight design of truss structures.

The surface shot peening coverage of industrial parts will affect the performance of the parts. Therefore, it is of great significance for the identification of the metal surface shot peening coverage area and the uncovered area for the metallurgical process. This paper uses the idea of image semantic segmentation, using CCD industrial camera to select and shoot the metal surface of 67%,89% and 98% shot peening coverage. Based on UNet and UNet++ model, the shot peening area is segmented. After training 200 epoch, the network converges and the training accuracy can reach 98.45%. Comparing the two kinds of network training results, using the UNet++ model with good results to predict the selected 18 metal surfaces, the proportion of shot peening coverage area and uncovered area is counted, and then the K-means clustering algorithm is used to classify. The classification accuracy can reach 100%.

Surround depth estimation provides a cost-effective alternative to LiDAR for 3D perception in autonomous driving. While recent self-supervised methods explore multi-camera settings to improve scale awareness and scene coverage, they are primarily designed for passenger vehicles and rarely consider articulated vehicles or robotics platforms. The articulated structure introduces complex cross-segment geometry and motion coupling, making consistent depth reasoning across views more challenging. In this work, we propose \\textbf{ArticuSurDepth}, a self-supervised framework for surround-view depth estimation on articulated vehicles that enhances depth learning through cross-view and cross-vehicle geometric consistency guided by structural priors from vision foundation model. Specifically, we introduce multi-view spatial context enrichment strategy and a cross-view surface normal constraint to improve structural coherence across spatial and temporal contexts. We further incorporate camera height regularization with ground plane-awareness to encourage metric depth estimation, together with cross-vehicle pose consistency that bridges motion estimation between articulated segments. To validate our proposed method, an articulated vehicle experiment platform was established with a dataset collected over it. Experiment results demonstrate state-of-the-art (SoTA) performance of depth estimation on our self-collected dataset as well as on DDAD, nuScenes, and KITTI benchmarks.

Drivers' visual attention provides critical cues for anticipating latent hazards and directly shapes decision-making and control maneuvers, where its absence can compromise traffic safety. To emulate drivers' perception patterns and advance visual attention prediction for intelligent vehicles, we propose DiffAttn, a diffusion-based framework that formulates this task as a conditional diffusion-denoising process, enabling more accurate modeling of drivers' attention. To capture both local and global scene features, we adopt Swin Transformer as encoder and design a decoder that combines a Feature Fusion Pyramid for cross-layer interaction with dense, multi-scale conditional diffusion to jointly enhance denoising learning and model fine-grained local and global scene contexts. Additionally, a large language model (LLM) layer is incorporated to enhance top-down semantic reasoning and improve sensitivity to safety-critical cues. Extensive experiments on four public datasets demonstrate that DiffAttn achieves state-of-the-art (SoTA) performance, surpassing most video-based, top-down-feature-driven, and LLM-enhanced baselines. Our framework further supports interpretable driver-centric scene understanding and has the potential to improve in-cabin human-machine interaction, risk perception, and drivers' state measurement in intelligent vehicles.

Background Dihydromyricetin (DHM), a flavonoid compound of natural origin, has been identified in high concentrations in ampelopsis grossedentata and has a broad spectrum of biological and pharmacological functions, particularly in regulating glucose and lipid metabolism. The objective of this research was to examine how DHM affected nonalcoholic fatty liver disease (NAFLD) and its underlying mechanisms involved in the progression of NAFLD in a rat model subjected to a high-fat diet (HFD). Additionally, the study examines the underlying mechanisms in a cellular model of steatohepatitis using palmitic acid (PA)-treated HepG2 cells, with a focus on the potential correlation between autophagy and hepatic insulin resistance (IR) in the progress of NAFLD. Methods SD rats were exposed to a HFD for a period of eight weeks, followed by a treatment with DHM (at doses of 50, 100, and 200 mg·kg −1 ·d −1 ) for additional six weeks. The HepG2 cells received a 0.5 mM PA treatment for 24 h, either alone or in conjunction with DHM (10 µM). The histopathological alterations were assessed by the use of Hematoxylin–eosin (H&E) staining. The quantification of glycogen content and lipid buildup in the liver was conducted by the use of PAS and Oil Red O staining techniques. Serum lipid and liver enzyme levels were also measured. Autophagic vesicle and autolysosome morphology was studied using electron microscopy. RT-qPCR and/or western blotting techniques were used to measure IR- and autophagy-related factors levels. Results The administration of DHM demonstrated efficacy in ameliorating hepatic steatosis, as seen in both in vivo and in vitro experimental models. Moreover, DHM administration significantly increased GLUT2 expression, decreased G6Pase and PEPCK expression, and improved IR in the hepatic tissue of rats fed a HFD and in cells exhibiting steatosis. DHM treatment elevated Beclin 1, ATG 5, and LC3-II levels in hepatic steatosis models, correlating with autolysosome formation. The expression of AMPK levels and its downstream target PGC-1α, and PPARα were decreased in HFD-fed rats and PA-treated hepatocytes, which were reversed through DHM treatment. AMPK/ PGC-1α and PPARα knockdown reduced the impact of DHM on hepatic autophagy, IR and accumulation of hepatic lipid. Conclusions Our findings revealed that AMPK/ PGC-1α, PPARα-dependent autophagy pathways in the pathophysiology of IR and hepatic steatosis has been shown, suggesting that DHM might potentially serve as a promising treatment option for addressing this disease. Graphical Abstract

Evaluation of nasal spray drug absorption has been challenging because deposited particles are consistently transported away by mucociliary clearance during diffusing through the mucus layer. This study developed a novel approach combining Computational Fluid Dynamics (CFD) techniques with a 1-D mucus diffusion model to better predict nasal spray drug absorption. This integrated CFD-diffusion approach comprised a preliminary simulation of nasal airflow, spray particle injection, followed by analysis of mucociliary clearance and drug solute diffusion through the mucus layer. The spray particle deposition distribution was validated experimentally and numerically, and the mucus velocity field was validated by comparing with previous studies. Total and regional drug absorption for solute radius in the range of 1 − 110 nm were investigated. The total drug absorption contributed by the spray particle deposition was calculated. The absorption contribution from particles that deposited on the anterior region was found to increase significantly as the solute radius became larger (diffusion became slower). This was because the particles were consistently moved out of the anterior region, and the delayed absorption ensured more solute to be absorbed by the posterior regions covered with respiratory epithelium. Future improvements in the spray drug absorption model were discussed. The results of this study are aimed at working towards a CFD-based integrated model for evaluating nasal spray bioequivalence.