Search Results (3,102)

Accurate object detection remains a major challenge in autonomous systems and surveillance, particularly when objects are partially or fully obscured by occlusions. To address this issue, we revisit FuDensityNet as a multimodal detection framework that jointly leverages 2D RGB images and 3D LiDAR point clouds for robust feature representation. The model integrates spatial and depth cues through low-rank tensor fusion (LRTF) and incorporates an Occlusion Rate (OR) assessment module that estimates the degree of occlusion and dynamically selects the most suitable detection pathway to preserve performance. Experiments on the KITTI and NuScenes datasets indicate that this adaptive strategy improves robustness under high occlusion while maintaining competitive accuracy in less challenging conditions. In particular, FuDensityNet attains 76.6% AP for car detection under “Hard” conditions on KITTI and outperforms several RGB-only and RGB–LiDAR baselines. Owing to its adaptive and modular design, FuDensityNet remains compatible with both 2D and 3D detection pipelines, making it a practical option for real-world environments where visual obstructions are frequent. Full article

Following the PRISMA-ScR framework for scoping reviews, we systematically searched five databases (Scopus, IEEE Xplore, ScienceDirect, SpringerLink, Web of Science) using a Boolean query combining real-time processing, 3D reconstruction, and deformation modelling terms. From 86 records identified, 56 peer-reviewed publications (2009–2025) were retained after two-stage screening and organized into a unified taxonomy covering sensing modalities (RGB-D, LiDAR, tactile), reconstruction pipelines (volumetric fusion, NRSfM, neural radiance fields), and deformation models (FEM, PBD, mass-spring, GNN-based surrogates, differentiable simulators). Of the 56 included works, 60% were published between 2022 and 2025, confirming the field’s rapid growth. Neural and implicit representations account for 20% of contributions, FEM-based methods for 16%, and hybrid or application-specific pipelines for 21%. Four systemic gaps emerge: the absence of a unified physics-aware benchmark; unresolved speed–accuracy trade-offs (PBD achieves >30 FPS on desktop GPUs for 10³–10⁴ vertex meshes but lacks mapping to physical material constants (Young’s modulus, Poisson’s ratio), limiting material fidelity; full-order FEM ensures physically consistent stress–strain behavior but runs at only 1–10 FPS without order reduction; reduced-order FEM recovers interactive rates for low-frequency deformation modes); fragile handling of occlusions and multi-object contact; and limited end-to-end integration of sensing and simulation. The findings support the presentation of a research roadmap centered on model order reduction, differentiable physics, multimodal sensing fusion, and standardized evaluation protocols, with implications for robust digital twins of deformable environments. Full article

Optical performance monitoring is essential for adaptive and intelligent coherent optical communication systems. In this paper, a Transformer-based multi-task meta-learning framework is proposed for joint modulation format identification and electrical signal-to-noise ratio (ESNR) estimation from original received waveforms. A simulated coherent optical communication system is established to generate QPSK, 16QAM, and 32QAM signals under different launch-power conditions. The received I/Q waveforms are directly used as model inputs, avoiding handcrafted feature extraction or constellation-image conversion. The proposed model employs a shared one-dimensional Transformer encoder to extract temporal waveform representations. A prototypical classification branch is used for few-shot modulation format identification, while an ESNR regression branch is introduced for continuous signal-quality estimation. The two tasks are jointly optimized under an episodic support-query training mechanism. Experimental results show that the proposed method achieves 99.99% modulation identification accuracy on the test episodes. For ESNR estimation, the model obtains an MAE of 0.1194 dB, an RMSE of 0.1738 dB, and an R² value of 99.83%. These results demonstrate that the proposed framework can simultaneously provide accurate modulation decisions and reliable ESNR estimation, showing its potential for waveform-based optical performance monitoring. Full article

Advances in imaging and fabrication technologies offer new opportunities to develop tools that support the visualization and understanding of complex biological materials. This contribution presents a comprehensive methodological framework for generating anatomically representative, species-specific 3D models of wood microstructure, intended to enhance student comprehension in wood science and related fields. The workflow integrates micro-X-ray computed tomography (micro-CT) scanning, image segmentation, STL model preparation, and additive manufacturing. Using micro-CT, we captured high-resolution, non-destructive 3D datasets of four wood species—European beech (Fagus sylvatica), oak (Quercus robur L.), Norway spruce (Picea abies), and Scots pine (Pinus sylvestris). The resulting volumetric data were processed with dedicated software to isolate and reconstruct key anatomical features, which were subsequently converted into printable STL models. These models were fabricated at a 1:400 scale using filaments composed of 40% wood particles and 60% biodegradable polylactic acid (PLA), underscoring the relevance of sustainable materials in educational tool development. The primary aim of this work is to document and justify each stage of the technological process, thereby providing a replicable pathway for producing detailed, pedagogically useful representations of wood microstructure. The resulting models are publicly available on the Sketchfab platform as part of the “3D Wood Micro Structure Collection.” Full article

In precision livestock farming, 3D point clouds provide important data support for analyzing pig behavior and monitoring their health. However, due to environmental occlusions, limited sensor viewpoints, and mutual shielding between pigs, the acquired point clouds are often severely partial, which affects the accuracy of body shape modeling and behavior recognition. To address these challenges, this study constructed a pig pose point cloud dataset using multi-view depth camera acquisition and point cloud registration techniques. Based on this dataset, an improved point cloud completion model, IIR-PoinTr, is proposed to enhance the reconstruction of geometric and topological structures in pig bodies. By strengthening local geometric perception and high-dimensional feature representation, the model improves the reconstruction quality of partial pig point clouds and produces more structurally consistent pig body shapes. Experimental results show that, on the self-constructed pig posture dataset, the proposed method reduces Chamfer Distance (CD-L1) by 3.6%, CD-L2 by 6.9%, and Earth Mover’s Distance (EMD) by 2.0%, while improving the F-score by 5.4% compared with the baseline model. In single-view point cloud completion tasks, the method is capable of reconstructing geometrically consistent pig body structures and increases downstream classification accuracy by 34.9%. These results indicate that the proposed method can improve the reconstruction quality of partial pig point clouds and provide preliminary technical support for posture analysis under occlusion. Full article

Incremental automatic modulation recognition is essential for the awareness of complex electromagnetic environments but is prone to catastrophic forgetting. This is fundamentally precipitated by shortcut learning, a phenomenon where deep models prioritize stable but non-essential channel artifacts (e.g., noise, fading) over intrinsic modulation characteristics. Consequently, models rely on spurious correlations that collapse during incremental task updates or environmental shifts, leading to representation drift. To bridge this gap, we propose the ParalIMR framework, which integrates a parallel reconstruction architecture with the segment substitution (SS) strategy to decouple modulation signatures from environmental fingerprints. Specifically, the parallel branch utilizes a Denoising AutoEncoder (DAE) as a task-agnostic structural anchor, purifying feature representations and maintaining geometric consistency across varying signal-to-noise ratios without propagating noise-overfitting to the classifier. In the meantime, the SS strategy actively disrupts the temporal coupling between class labels and hardware fingerprints through random reorganization, forcing the model to extract modulation-invariant structural cues. Experimental results on the RML2016a datasets demonstrate that in a three-stage incremental setup, our method achieves an overall accuracy of 84.32% at 0 dB SNR, representing a 2.69% improvement over the iCaRL baseline. Notably, this advantage expanded to 4.46% on RML2018, demonstrating that ParalIMR effectively arrests catastrophic forgetting. Ultimately, this research provides a robust learning paradigm tailored for cognitive radio and electronic warfare in dynamic electromagnetic landscapes. Full article

Multimodal recommender systems are challenged by heterogeneous modality noise and coarse-grained feature fusion. Specifically, existing frequency-domain methods typically apply symmetric filtering across modalities, ignoring their distinct spectral characteristics. Consequently, symmetric filtering cannot simultaneously satisfy the denoising requirements of visual features and the semantic preservation requirements of textual features, leading to suboptimal multimodal representations. Meanwhile, current fusion strategies mainly operate at the instance level with static modality weights, lacking flexibility to dynamically adjust feature channels for user-specific collaborative contexts. To address these issues, this paper proposes MFA-GCN, a multimodal recommendation framework that combines asymmetric spectral filtering, multiview graph enhancement, and behavior-guided channel attention. For visual modalities, a multiscale frequency-domain module integrating 1D convolution and self-attention is adopted to suppress high-frequency disturbances while preserving informative structures. For textual modalities, a lightweight complex-domain scaling strategy is introduced to adjust spectral energy while maintaining semantic consistency. In addition, auxiliary user–user and item–item graphs are constructed to supplement sparse user–item interactions and provide richer collaborative signals. A behavior-guided channel attention mechanism is further used to dynamically refine multimodal representations. Experiments on three public Amazon datasets demonstrate that MFA-GCN consistently outperforms several representative baselines. Full article

Background/Objectives: Statistical shape modeling (SSM) is used to describe transtibial residual-limb morphology for prosthetic socket design, simulation, and future structural testing. However, conventional intrinsic metrics such as compactness, generality, and specificity may not directly reflect geometric fidelity to the original shape. This study examined the relationship between geometric fidelity and SSM evaluation and assessed a cross-sectional shape representation framework for transtibial residual limbs. Methods: Residual-limb surfaces were acquired from 62 adults with unilateral transtibial amputation using a structured-light 3D scanner while preserving habitual limb posture. Two surface-based registration methods, non-rigid iterative closest point and Bayesian coherent point drift, were compared with a cross-sectional representation in which proximal and distal regions were sectioned separately and reconstructed by strip triangulation. Geometric fidelity to the original mesh was quantified using average symmetric surface distance (ASSD). SSM performance was evaluated using compactness, generality, and specificity. Results: The optimal cross-sectional configuration was 60 sections × 72 points. The proposed method showed the best geometric fidelity (ASSD, 1.30 ± 0.14 mm), followed by Bayesian coherent point drift (1.33 ± 0.14 mm) and non-rigid iterative closest point (1.48 ± 0.48 mm). Compactness was highest for the proposed method, reaching 95% cumulative variance in four modes, compared with five and seven modes, respectively, for the two surface-based methods. In geometry-space evaluation, the proposed method showed the lowest specificity error, while differences in generality were statistically significant but small in magnitude. Conclusions: Intrinsic SSM metrics alone were insufficient to judge registration quality in transtibial residual-limb modeling. The cross-sectional representation preserved the original surface geometry more faithfully than the evaluated surface-based methods while maintaining competitive SSM performance. Full article

Background: MS/MS spectral databases provide reference spectra for compound identification in metabolomics studies. Their utility depends on clear links among compound identifiers, chemical structures, and MS/MS evidence, yet these links are often complicated by database-specific identifiers, heterogeneous structural representations, and stereochemical specifications. Methods: Here, we present a two-layer structural key framework for linking compound identifiers and MS/MS evidence through standardized structures. Reported SMILES were standardized and converted into InChIKey-derived stereo keys and connectivity keys using a Python-based RDKit workflow. Results: As illustrated using stereoisomeric cases such as L- and D-proline, the stereo key layer preserves compound identifiers and metadata at the stereo level, whereas the connectivity key layer groups comparable MS/MS evidence at the molecular connectivity level. In a database-scale application, 217,920 HMDB compound entries were organized into 216,783 stereo keys and 196,512 connectivity keys, and 144,591 spectra from the spectrum-centered MoNA database were incorporated into the HMDB-centered framework, increasing MS/MS evidence coverage, particularly at the molecular connectivity level. Conclusions: Together, this framework links compound identifiers, standardized structures, and MS/MS evidence at the stereo and connectivity levels, providing a bidirectionally traceable system for spectral database curation without forcing connectivity-level MS/MS evidence into stereo-specific compound identities. Full article

Sublobar pulmonary resection has become an increasingly adopted approach for early-stage non-small cell lung cancer, driven by evidence that anatomical segmentectomy can achieve oncological outcomes comparable to lobectomy in selected patients. Safe execution of sublobar resection depends on accurate preoperative identification of segmental bronchovascular anatomy, which demonstrates substantial variability. Conventional two-dimensional (2D) computed tomography (CT) imposes significant limitations on anatomical interpretation, particularly at the segmental and subsegmental level. Three-dimensional (3D) bronchovascular modelling provides patient-specific representations of segmental anatomy and relationships that address these limitations. This narrative review examines the current and emerging roles of 3D modelling in personalised thoracic surgery. It discusses the anatomical basis for its application, the limitations of conventional imaging, and the contribution of 3D modelling to preoperative planning and intraoperative decision making. It also considers broader applications, current limitations, and future directions, with emphasis on how patient-specific 3D modelling can support more tailored operative strategies and more individualised surgical care. Full article

This paper presents GenPluSSS, a plugin that adds the visualization of homogeneous and heterogeneous, optically thick, translucent materials on the Blender 3D modeling tool. The working principle of this plugin is based on the GenSSS method, which combines Genetic Algorithm (GA) and Singular Value Decomposition (SVD)-based subsurface scattering representation. The proposed plugin has been implemented using the Mitsuba renderer, an open-source rendering system, and has been validated on measured subsurface scattering datasets. Experimental results demonstrate that the proposed plugin visualizes homogeneous and heterogeneous subsurface scattering effects accurately with compact data representation while maintaining computational efficiency and achieving competitive rendering times compared to dipole-based and SVD-based approaches. In addition, conceptual and quantitative comparisons with recent neural subsurface scattering methods are presented in terms of rendering speed, peak memory usage, material support, and hardware dependency. The proposed framework brings measured subsurface scattering methods into practical rendering workflows within open-source content creation environments. Full article

It is well-known that activation functions are crucial in determining spectral expressiveness, training dynamics, and reconstruction accuracy in implicit neural representations (INRs), which employ coordinate-based multilayer perceptrons to represent continuous signals. Despite showing excellent performance, sinusoidal activations, for example SIREN, are limited in their adaptability to diverse signal types due to their fixed harmonic structure. In this paper, we propose two novel periodic activation functions for INRs. (1) Harmonic generalizes sinusoidal activations by combining the fundamental frequency with learned second and third harmonics through per-neuron trainable amplitude coefficients, resulting in a richer spectral basis within the SIREN initialization framework. (2) PM-FINER (Phase-Modulated FINER) extends the variable-periodic FINER activation by embedding frequency modulation synthesis directly into the instantaneous phase, enabling data-driven phase distortion via a learnable modulation index and carrier ratio. We conducted comprehensive experiments spanning nine architectural configurations (including SIREN, WIRE, FINER, Gaussian, Harmonic, PM-FINER, and an additional direct comparison against the Subtractive Modulative Network (SMN)), using six natural images, three learning rate schedulers, and three random seeds, totaling 486 main training runs (534 runs total including an ${\omega }_0$ sensitivity sweep). Our evaluation combined peak signal-to-noise ratio (PSNR), structural similarity index (SSIM), and rigorous statistical analysis, such as paired t-tests, Wilcoxon signed-rank tests, Cohen’s d effect sizes, and Friedman rank tests. Under cosine annealing, Harmonic achieves a mean PSNR gain of 6.08 dB over SIREN and 2.57 dB over FINER (both $p<0.001$, Cohen’s $d>3.7$), while PM-FINER ranks statistically on par with Harmonic (mean difference $0.17$ dB, $p=0.36$), outperforming all of the other baselines. Compared with SMN, Harmonic outperforms it by $+7.94$ dB under cosine annealing (Bonferroni-adjusted $p<{10}^{-5}$, Cohen’s $d=12.3$), winning on all six images. Additionally, the Friedman ranking across the six images confirmed Harmonic (with mean rank $=1.33$) and PM-FINER (with mean rank $=1.67$), being the top two methods under cosine annealing. Our results establish interpretable multi-harmonic and phase-modulated activations as real alternatives to the existing INR activation functions. Full article

Reliable fault diagnosis of rotating machinery is critical for averting serious failures in modern industrial systems. While data-driven deep learning has advanced condition monitoring, its success is fundamentally predicated on the availability of independent and identically distributed (I.I.D.) datasets. In realistic operational environments, machinery frequently experiences massive domain shifts induced by varying rotational speeds. Concurrently, acquiring high-fidelity fault instances is limited compared to abundant healthy baseline data, often resulting in a long-tailed distribution. Under such data-starved conditions, conventional few-shot domain adaptation (FSDA) methodologies often may be affected by distributional erasure; global alignment objectives are mainly driven by the healthy majority, causing sparse fault signatures to be erroneously absorbed as noise and leading to severe diagnostic performance degradation. To address this setting, this study develops a physics-guided dual-stream fusion framework for extreme few-shot cross-domain fault diagnosis. The method does not treat the Laplace wavelet, STFT, CNNs, or AdaBN as newly introduced techniques. Instead, it integrates these components into a unified diagnostic pipeline designed for long-tailed target support sets under large speed shifts. A learnable Laplace wavelet convolution is used in the temporal branch to emphasize transient impact responses, while STFT spectrograms provide a complementary time-frequency representation for the two-dimensional branch. The two feature streams are then fused for target fault classification. For domain adaptation, a Strict AdaBN strategy is applied using only the target support set, rather than the target test data or a large unlabeled target pool. Under the evaluated 50 healthy + 12 fault support condition, the healthy samples provide target-domain operating-background statistics for BN recalibration, while the limited fault samples are used for supervised classifier adjustment. Experiments on the HUSTbearing and Torino DIRG datasets show that the proposed integrated framework achieves stable performance under the evaluated few-shot cross-speed settings. These results suggest that combining physics-guided Laplace convolution, time-frequency representations, and support-set-restricted BN recalibration can be useful for bearing fault diagnosis when target fault samples are limited. Full article

Superheavy and hyperheavy nuclei are one of the frontiers of nuclear structure nowadays, while for many actinides rather limited experimental information exists. Therefore, theoretical methods providing parameter-independent predictions for these nuclei are of particular interest. Such a method is the proxy-SU(3) approximation to the shell model, which has been adequately tested against experimental data in medium-mass and heavy nuclei up to the rare-earth region, and it has been found to provide reliable, parameter-independent predictions for the collective deformation variables $\beta$ and $\gamma$. Within the proxy-SU(3) approach, the SU(3) symmetry of the three-dimensional harmonic oscillator, which is destroyed beyond the $sd$ shell by the strong spin–orbit interaction, is restored through a unitary transformation. For each nucleus, the most symmetric irreducible representation (irrep) allowed by the Pauli principle and the short-range nature of the nucleon–nucleon interaction, called the highest-weight (hw) irrep in mathematical language, is found to suffice, except in cases in which the hw irrep turns out to be completely symmetric, so that the next highest weight (nhw) irrep has also to be included. In this article we provide a full collection of the hw and nhw irreps, as well as of the corresponding parameter-free predictions for the deformation variables $\beta$ and $\gamma$, for all atomic nuclei ranging from $Z=82$, $N=82$ to $Z=126$, $N=258$. Several cases exemplifying the use of the collected results for studying the prolate-to-oblate shape transition, mirror symmetries, and the evolution of the collective variables along the valley of stability are also considered. Full article

Face super-resolution (FSR) aims to reconstruct high-quality high-resolution face images from low-resolution inputs. Although CNN–Transformer hybrid models have shown promising performance by jointly modeling local textures and global dependencies, their large parameter sizes and high computational costs hinder practical deployment in resource-constrained scenarios such as mobile devices and embedded systems. Meanwhile, existing lightweight SR models usually reduce complexity by simplifying network depth, channel dimensions, or convolutional operations, which may weaken feature representation capability and lead to insufficient recovery of fine facial structures. To address these issues, this paper proposes HCTIUNet, a lightweight CNN–Transformer hybrid network based on an inverted U-shaped architecture. Specifically, the proposed network integrates lightweight CNN branches for local facial texture extraction and Transformer branches for global dependency modeling, while introducing a multi-scale feature interaction strategy and a global feature refinement module to enhance facial structural details. Experimental results on the FFHQ, CelebA, and Helen datasets demonstrate that HCTIUNet achieves competitive performance under the ×8 face super-resolution setting, obtaining PSNR/SSIM/LPIPS values of 27.55 dB/0.765/0.225, 27.63 dB/0.761/0.212, and 27.53 dB/0.777/0.213, respectively. Moreover, HCTIUNet contains 10.5 M parameters, requires 9.9 G FLOPs, and achieves an inference time of 0.021 s. These results indicate that the proposed method achieves a favorable trade-off between reconstruction accuracy, perceptual quality, and computational efficiency, making it suitable for efficient face super-resolution applications. Full article