Traffic safety is a critical issue in urban transportation systems, with crash severity predictions playing a key role in mitigating road traffic injuries and fatalities. This study addresses the research question: how can transforming tabular crash data into image-based representations using deep learning techniques improve the prediction of crash severity, while effectively managing class imbalances? For analysis, 63,745 barrier crashes that occurred in Texas from 2017–2022 were examined. Using a sample database of barrier crash incidents, we introduce a novel approach that leverages Convolutional Neural Networks (CNNs) to convert traditional tabular crash data into image representations by employing the DeepInsight technique, which maps feature similarity into structured two-dimensional grids, enhancing the accuracy and effectiveness of prediction models. Generative Adversarial Networks (GANs) were used for data augmentation, addressing issues of class imbalances with innovative sampling techniques including under-sampling, over-sampling, and balanced sampling. Key findings indicate that balanced datasets significantly improve predictive performance, enhancing generalizability across diverse crash scenarios. Moreover, under- and over-sampling strategies offer a refined trade-off among precision, recall, and F1-score, while the over-sampling method distinctly excels in practical model performance by reducing misclassification rates and improving overall accuracy. Comparative evaluation against an XGBoost baseline on identical balanced data highlights the ability of the image-based deep learning framework to capture complex crash patterns. These findings suggest that integrating feature-similarity–based tabular-to-image transformation with deep learning and balanced data sampling can significantly improve crash severity classification.

The height of a binary sequence serves as a metric that captures the randomness of a sequence. In this work, the height of a finite binary sequence, which was a missing random variable in the recent work published at SETA 2024, is defined, and an explicit formula is derived for the probability values of a sequence of length <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i>, to have a height equal to <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">t</i>, in terms of binomial coefficients. While this formula yields useful insights, computing exact values of binomial coefficients and those probability values becomes infeasible as <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">n</i> gets larger. A method to overcome this problem is to use asymptotic approaches, which cause some errors when used in randomness tests. In this work, a recursive formula is derived for the same probability values, and the exact probability values for sequences of length up to at least 4096 is obtained. Using these exact values in a random walk randomness test, one can obtain more accurate results. In addition, formulas for the expected value and variance of the random variable that assigns the number of balanced points in a binary sequence of length n were derived, which had not been addressed in the previous work published at SETA 2024. A new randomness test is proposed based on random walk height.

Ground roll is a dominant coherent noise in land seismic data, characterized by low frequency, low velocity, and high amplitude. often overlaps with reflection arrivals, thereby reducing the reliability of seismic imaging and interpretation. Existing deep learning approaches exhibit inherent limitations in ground-roll attenuation, since Convolutional Neural Network-based models are constrained by local receptive fields, while standard Transformer architectures require large training datasets and incur substantial computational and memory costs due to their global self-attention mechanism. In this study, we propose a U-shaped Swin Transformer (U-Swin Transformer) for ground-roll attenuation. The proposed model replaces global self-attention with window-based self-attention, achieving linear computational complexity with respect to the input size and thereby significantly improving efficiency and scalability. Two near-surface field datasets acquired from the inland sabkha of Jayb Uwayyid (SJUW) in the Saudi Kingdom and the coastal sabkha of Abu Dhabi (CSAD) are used to evaluate the method under realistic geological conditions. Experimental results demonstrate that the proposed U-Swin Transformer effectively handles aliased ground roll and signal-noise overlap, which remain challenging for the U-shaped neural network (U-Net) and F-K filtering. Although the denoising performance of the U-Swin Transformer is comparable to that of the Vision Transformer (ViT) with similar mean absolute error, peak signal-to-noise ratio, and structural similarity index measure, it achieves this performance with significantly lower computational cost and GPU memory consumption. Furthermore, incorporating a transfer learning strategy enables efficient adaptation of the U-Swin Transformer across different sabkha seismic environments, substantially reducing training time compared to training the model from scratch.

This paper presents a statistical framework for characterizing quantum noise effects in amplitude-encoded images through the analysis of noise-induced amplitude perturbations. Classical grayscale images are encoded as normalized quantum amplitude vectors and subjected to stochastic amplitude damping, phase damping, and depolarizing channels. The resulting perturbations are quantified and modeled using a bounded-domain distribution obtained via a nonlinear transformation of the Ola distribution. A single shape parameter is estimated via maximum-likelihood methods, providing a compact descriptor of perturbation concentration across images, noise channels, and noise levels. Experimental results reveal a consistent decreasing relationship between the estimated parameter and noise level, along with clear channel-dependent statistical signatures. Phase damping, despite producing negligible visual changes in reconstructed images, yields measurable perturbations that are accurately captured by the proposed model. Goodness-of-fit analysis based on Kullback–Leibler divergence, mean absolute error, and likelihood-based criteria demonstrates stable model performance across noise regimes, supported by bootstrap uncertainty analysis. The proposed approach provides a unified and statistically robust representation of quantum noise-induced perturbations in amplitude-encoded images, enabling systematic comparison of noise effects beyond visual inspection. This characterization supports noise-aware quantum image processing by offering a compact, channel-dependent statistical description that can inform algorithm design and evaluation under realistic noise conditions.

BLDC motors are increasingly used in battery-powered and industrial applications due to their high efficiency and reliability. However, their performance is often degraded by current ripple and torque pulsations caused by higher-order harmonic components in the back electromotive force, which are not sufficiently addressed by conventional control strategies such as field-oriented control. This paper presents a feed-forward method for improving BLDC motor efficiency by injecting selected higher time-harmonic components into the inverter output voltages. The method aims to reduce current ripple and optimize torque generation without requiring additional hardware or significant modifications to standard FOC algorithms, making it suitable for existing motor control platforms. Simulation and experimental validation confirm that the proposed approach decreases phase current RMS values and minimizes torque pulsations. Experimental validation on a low-voltage BLDC motor shows that overall efficiency improvements are most significant at high speeds and low loads, with a maximum measured gain of 0.5%. Due to its simplicity and low computational requirements, the method is well-suited for embedded implementations in cost-sensitive applications.

Hydraulic systems are widely used in various applications, including industrial production, off-road mobile machinery, and aerospace. The failure of a component within a hydraulic system can lead to significant damage and maintenance downtime. To prevent such issues, automated condition monitoring techniques are emerging, relying on data-driven methods such as machine learning algorithms. However, most existing solutions depend on system and application-specific data, which can be costly to obtain in real-world scenarios. Despite this, different hydraulic systems often share structural similarities and common physical mechanisms that lead to faults and failures. This raises the question of whether these similarities can be leveraged to develop more efficient data-driven condition monitoring solutions for hydraulic applications. Therefore, this paper focuses on developing an automated condition monitoring solution that can be independently applied across similar hydraulic systems using transfer learning. The systems considered in this study are a hydraulic press and an electro-hydraulic reach truck. For both systems, simulation data is generated, including scenarios of healthy operation and six different fault conditions. A data-centric transfer approach is pursued by applying the domain adaptation algorithm called Transfer Adaptive Boosting (TrAdaBoost). Additionally, cases with varying degrees of fault complexity and different transfer directions are evaluated. The results demonstrate that the proposed framework is suitable for process independent condition monitoring with a reasonable average transfer accuracy of 93.6% for the presence of non-concurrent faults and 85.7% with the presence of concurrent faults, representing a substantial improvement compared to the baseline transfer results, which remain near 50% when domain adaptation is not applied.

The Segment Anything Model (SAM) has demonstrated impressive generalization in prompt-based segmentation. Yet, the potential of semantic text prompts remains underexplored compared to traditional spatial prompts, such as points and boxes. This paper introduces SAM-PTx, a parameter-efficient approach for adapting SAM using frozen CLIP-derived text embeddings as class-level semantic guidance. Specifically, we propose a lightweight adapter, the Parallel-Text Adapter, that injects text embeddings into SAM’s image encoder, enabling semantics-guided segmentation while keeping most of the original architecture frozen. Through supervised experiments and ablations on the low-data subsets of COCO and ADE20K as well as the COD10K dataset, we show that incorporating fixed text embeddings as additional inputs improves segmentation performance over purely spatial prompt baselines. To our knowledge, this is the first work to use text prompts for segmentation on the COD10K dataset. These results suggest that integrating semantic conditioning into SAM’s architecture offers a practical, scalable path to efficient adaptation with minimal computational complexity.

The estimation of 6D pose from RGB-D data remains challenging due to occlusions, textureless objects, and depth noise. In this work, We introduce a novel architecture to calculate precisely the 6DoF object pose from a single RGB-D image. Unlike existing structures that rely on direct regression & convolution based pose estimation as well as heavily depend on large model training, our vision based dual stream approach addresses this challenging task using hybrid multi modal fusion architecture combining self-supervised vision transformers (DINOv2) and attention based point cloud processing using C3G (Compact 3D Gaussian representations integrated with Point Transformer V3). The DINOv2 approach provides robust semantic understanding without requiring fine-tuning of visual backbone, while Point Transformer V3 employs vector attention mechanisms to model complex 3D geometric patterns from depth point clouds. Moreover, we present a mask-guided point cloud extraction approach that concentrates processing on object relevant regions while filtering out background noise. The model’s efficacy is demonstrated by the experimental results on the LineMOD-Occluded dataset over RDPN SOTA benchmark, which show that our network requires substantially fewer trainable parameters than fully-supervised alternatives while achieving competitive performance and notable improvements with ADD(S) metric, rotation error, and translation error. Self-supervised learning and attention based geometric reasoning together provide new era for data efficient 6D pose estimation.