Nuctech (China)

This study proposes a fiber optic sensing and multimodal decision-making based security system to enhance nuclear facility perimeter protection in extreme environments. The system integrates a specially designed optical cable into a metal fence structure, combining physical barrier and intrusion detection. To address environmental interference, a Multimodal Decision Algorithm with Dynamic Deformation Threshold (MDTA) is developed, implementing a three-level processing mechanism: adaptive Kalman filtering for wind vibration noise reduction, dual-wavelength detection for temperature compensation, and health assessment matrix generation. A gradient-weighted fusion strategy dynamically optimizes thresholds for various intrusion types. Experimental results demonstrate outstanding performance: <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{9 8. 7 \%}$</tex> intrusion recognition accuracy, 0.03 false alarms/year under level 12 winds, and 0.8 -second response time (37 % improvement over industry standards). The system maintains 2.4 % detection interference in heavy rain and achieves <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\pm 1.5 ~\mathrm{m}$</tex> positioning accuracy across full environmental conditions (<tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathbf{4 0} \boldsymbol{\sim} \mathbf{8 0} \boldsymbol{\circ}^{\circ} \mathrm{C} \boldsymbol{,} \mathbf{1 0} \boldsymbol{\sim} \mathbf{9 5 \%} \mathbf{R H} \boldsymbol{)}$</tex>. Supported by domestic hardware, this solution provides highly reliable, robust protection for nuclear facilities.

This study proposes a deep learning-based approach for correcting stripe-like scattering artifacts in high-energy X-ray digital radiography (DR) systems for cargo and vehicle screening. These artifacts, mainly caused by internal system structures and detector cross-talk, degrade image quality. Instead of relying on traditional scattering kernel-based corrections with limited adaptability, the method learns the spatial variation pattern of system-induced artifacts. Scattering-dominant images are acquired using high-density shielding to suppress primary transmission and highlight system artifacts. A lightweight convolutional neural network is trained to capture normalized spatial distribution patterns of the streaks. During inference, the input image's amplitude is convolved with the learned pattern to estimate the scatter component, which is then subtracted or compensated for accordingly. Experiments on real-world DR systems confirm effective suppression of system-generated stripes and improved image clarity, with strong generalizability across objects.

Isocyanate-based polyimide (IBPI) films synthesized using isocyanates and anhydrides as reactive monomers exhibit inferior mechanical performance, limiting their practical applicability. This study introduced 4,4′-oxidianiline (ODA) as a “bridging agent” to connect anhydride-terminated oligomer chain segments generated from diphenylmethane diisocyanate (MDI-50) and 4,4′-oxydiphthalic dianhydride (ODPA). A high-performance IBPI film with high solubility in common organic solvents was obtained via a bridging strategy; this work marked it as PI-X. The mechanical, thermal, and electrical properties of the PI-1.0 film were significantly improved. The tensile strength increased from 54.97 to 108.87 MPa, representing a 98% increase over that of the pure IBPI film. The temperature at 5% weight loss increased from 461 to 524 °C. Meanwhile, the dielectric constant and loss tangent decreased to 2.21 and 0.0047 at 1 kHz, respectively. These excellent performances show that IBPI films have great application prospects in the microelectronics industry.

Heart disease is the world’s leading cause of death. Diagnostic models based on electrocardiograms (ECGs) are often limited by the scarcity of high-quality data and issues of data imbalance. To address these challenges, we propose a conditional generative adversarial network (CECG-GAN). This strategy enables the generation of samples that closely approximate the distribution of ECG data. Additionally, CECG-GAN addresses waveform jitter, slow processing speeds, and dataset imbalance issues through the integration of a transformer architecture. We evaluated this approach using two datasets: MIT-BIH and CSPC2020. The experimental results demonstrate that CECG-GAN achieves outstanding performance metrics. Notably, the percentage root mean square difference (PRD) reached 55.048, indicating a high degree of similarity between generated and actual ECG waveforms. Additionally, the Fréchet distance (FD) was approximately 1.139, the root mean square error (RMSE) registered at 0.232, and the mean absolute error (MAE) was recorded at 0.166.

Background:Epilepsy is a common neurological disorder that affects approximately 60 million people worldwide. Characterized by unpredictable neural electrical activity abnormalities, it results in seizures with varying intensity levels. Electroencephalography (EEG), as a crucial technology for monitoring and predicting epileptic seizures, plays an essential role in improving the quality of life for people with epilepsy.Method:This study introduces an innovative deep learning model, a lightweight triscale yielding convolutional neural network” (LTY-CNN), that is specifically designed for EEG signal analysis. The model integrates a parallel convolutional structure with a multihead attention mechanism to capture complex EEG signal features across multiple scales and enhance the efficiency achieved when processing time series data. The lightweight design of the LTY-CNN enables it to maintain high performance in environments with limited computational resources while preserving the interpretability and maintainability of the model.Results:In tests conducted on the SWEC-ETHZ and CHB-MIT datasets, the LTY-CNN demonstrated outstanding performance. On the SWEC-ETHZ dataset, the LTY-CNN achieved an accuracy of 99.9%, an area under the receiver operating characteristic curve (AUROC) of 0.99, a sensitivity of 99.9%, and a specificity of 98.8%. Furthermore, on the CHB-MIT dataset, it recorded an accuracy of 99%, an AUROC of 0.932, a sensitivity of 99.1%, and a specificity of 93.2%. These results signify the remarkable ability of the LTY-CNN to distinguish between epileptic seizures and nonseizure events. Compared to other existing epilepsy detection classifiers, the LTY-CNN attained higher accuracy and sensitivity.Conclusion:The high accuracy and sensitivity of the LTY-CNN model demonstrate its significant potential for epilepsy management, particularly in terms of predicting and mitigating epileptic seizures. Its value in personalized treatments and widespread clinical applications reflects the broad prospects of deep learning in the health care sector. This also highlights the crucial role of technological innovation in enhancing the quality of life experienced by patients.

Remote sensing (RS) multi-class object detection holds great potential in various fields such as agriculture production, urban planning and military reconnaissance. Recent research on RS object detection has primarily focused on utilizing multi-scale feature while overlooking the unique prior knowledge in global information. This prior knowledge can be valuable since small RS objects may be incorrectly detected without sufficient reference to global information, and different types of objects may require different global information. In this paper, we address this prior factor by introducing recursive gate-controlled convolution (RGC). RGC possesses the capability to model global information, allowing for the recursive detection of spatially adjacent features related to the objects through simple operations. By doing so, it better emulates the detection performance of various objects in RS scenario. Additionally, recognizing the crucial role played by multi-scale features in dense scenes, we have preserved this knowledge and enhanced the interaction between features using the bidirectional feature pyramid. Our experimental results on the FAIR1M have achieved competitive performance (68.0% mAP), significantly surpassing the methods employed in FAIR1M.

Purpose FLASH (ultra-high dose rate) radiotherapy spares normal tissue while keeping tumor control. However, the mechanism of the FLASH effect remains unclear and may have consequences beyond the irradiated area. Materials and methods We reanalyze the available results of ultra-high-dose-rate-related experiments to find out the key points of the mechanism of the FLASH effect. Then, we present a hypothesis on the mechanism of the FLASH effect: FLASH beams generate a high transient concentration of peroxyl radicals leading to a high fraction of radical recombination, which results in less oxidation damage to normal tissue. For the cells containing higher concentrations of antioxidants, the fractions of radical recombination are smaller because the antioxidants compete to react with peroxyl radicals. Therefore the damages by different dose rate beams differ slightly in this condition. Since some tumors contain a higher level of antioxidants, this may be the reason for the loss of the protective effect in tumors irradiated by FLASH beams. The high concentration of antioxidants in tumors results in slight radiolytic oxygen consumption, and consequently the protective effect observed in in vitro experiment cannot be observed in in vivo experiment. To quantitatively elaborate our hypothesis, a kinetic model is implemented to simulate the reactions induced by irradiation. Two parameters are defined to abstractly study the factors affecting the reaction, such as dose rate, antioxidants, total dose and reaction rate constants. Results and conclusions We find that the explanation of the difference between in vivo and in vitro experiments is crucial to understanding the mechanism of the FLASH effect. Our hypothesis agrees with the results of related experiments. Based on the kinetic model, the effects of these factors on the FLASH effect are quantitatively investigated.

The mechanism of the FLASH effect remains unclear and could be revealed by studying chemical reactions during irradiation. Monte Carlo simulation of the radiolytic species is an effective tool to analyze chemical reactions, but the simulation is limited by computing costs of the step-by-step simulation of radiolytic species, especially when considering beam with complex time structure. The complexity of the time structure of beams from accelerators in FLASH radiotherapy requires a high-performance Monte Carlo code. In this work, we develop a CPU-GPU coupling accelerating code with the independent reaction times (IRT) method to extend the chemical module of our nanodosimetry Monte Carlo code NASIC. Every chemical molecule in the microenvironment contains time information to consider the reactions from different tracks and simulate beams with complex time structures. Performance test shows that our code significantly improved the computing efficiency of the chemical module by four orders of magnitude. Then the code is used to study the oxygen depletion hypothesis in FLASH radiotherapy for different conditions by setting different parameters. The transient oxygen consumption rate values in the water are calculated when the pulses width ranges from 2 ps to 2 μs, the total dose ranges from 0.5 Gy to 100 Gy and the initial oxygen concentration ranges from 0.1% to 21%. The time evolution curves are simulated to study the effect of the time structure of an electron linear accelerator. Results show that the total dose in several microseconds is a better indicator reflecting the radiolytic oxygen consumption rate than the dose rate. The initial oxygen greatly affects the oxygen consumption rate because of the reaction competition. The diffusion of oxygen determined by the physiological parameters is the key factor affecting oxygen depletion during the radiation using electron linear accelerators. Our code provides an efficient tool for simulating water radiolysis in different conditions.

Research regarding cellular responses at different oxygen concentrations (OCs) is of immense interest within the field of radiobiology. Therefore, this study aimed to develop a mechanistic model to analyze cellular responses at different OCs. A DNA damage model (the different cell oxygen level DNA damage [DICOLDD] model) that examines the oxygen effect was developed based on the oxygen fixation hypothesis, which states that dissolved oxygen can modify the reaction kinetics of DNA-derived radicals generated by ionizing radiation. The generation of DNA-derived radicals was simulated using the Monte Carlo method. The decay of DNA-derived radicals due to the competing processes of chemical repair, oxygen fixation, and intrinsic damaging was described using differential equations. The DICOLDD model was fitted to the previous experimental data obtained under different irradiation configurations and validated by calculating the yields of DNA double-strand breaks (DSBs) after exposure to 137 Cs as well as cell survival fractions (SFs) using a mechanistic model of cellular survival. Moreover, we used the DICOLDD model to calculate DNA DSB damage yields after irradiation with 0.5-50MeV protons. Generally, DSB yields calculated after exposure to 137 Cs at different OCs correspond to statistical uncertainties of previous experimental results. Calculated SFs of CHO and V79 cells exposed to photons, protons, and alpha particles at different OCs generally concur with those obtained in previous studies. Our results demonstrated that the variation in DSB yields was less than 10% when the cellular OC decreased from 21% to 5%. Additionally, DSB yields changed drastically when OC dropped below 1%. We developed a DNA damage model to evaluate the oxygen effect and provide evidence that a reaction-kinetic model of DNA-derived radicals induced by ionizing radiation suffices to explain the observed oxygen effects. Therefore, the DICOLDD model is a powerful tool for the analysis of cellular responses at different OCs after exposure to different types of radiation.