Sensors

20 pages, 1861 KB

Open AccessArticle

Compensation Strategies in Post-Stroke Individuals: Insights from Upper Body Kinematics Analysis Based on Inertial Sensors

by Carrie-Louise Thouant, Elena Sofia Cocco, Giovanni Morone, Carlotta Maria Manzia, Francesco Infarinato, Paola Romano, Matteo Cioeta, Michela Goffredo, Marco Franceschini and Sanaz Pournajaf

Sensors 2025, 25(24), 7609; https://doi.org/10.3390/s25247609 (registering DOI) - 15 Dec 2025

Abstract

Background: One of the main goals of rehabilitation after stroke is the restoration of motor function. Understanding movement patterns and compensatory strategies is essential to optimize therapy. This study analyzes upper limb kinematics during the Box and Block Test (BBT) to identify and [...] Read more.

Background: One of the main goals of rehabilitation after stroke is the restoration of motor function. Understanding movement patterns and compensatory strategies is essential to optimize therapy. This study analyzes upper limb kinematics during the Box and Block Test (BBT) to identify and quantify typical post-stroke compensation strategies. Methods: Thirty-one sub-acute stroke participants and thirty-one healthy controls were included. Kinematic data were collected using a 7-IMU system. Joint angles were analyzed with MATLAB R2023a, and 3D trajectories were reconstructed from calibrated quaternions and anthropometric data. Group differences were assessed with the Mann–Whitney test. Compensation strategies were quantified in percentage terms relative to healthy subjects. Results: Significant intergroup differences were observed in mean joint angles and ranges of motion. On the paretic side, participants overused the wrist and shoulder to compensate for reduced elbow and trunk activity. Similar overuse was also observed on the unaffected side. Quantification showed that 83.9% and 80.6% compensate, respectively, with wrist and trunk and 67.7% with the shoulder. Conclusions: Using IMUs during the BBT, this study identified specific compensation strategies that may hinder recovery. It also contributed to developing a quantification scale, supporting more personalized rehabilitation and improved quality of life. Full article

(This article belongs to the Special Issue IMU and Innovative Sensors for Healthcare)

► Show Figures

Graphical abstract

21 pages, 920 KB

Open AccessArticle

Audio Deepfake Detection via a Fuzzy Dual-Path Time-Frequency Attention Network

by Jinzi Li, Hexu Wang, Fei Xie, Xiaozhou Feng, Jiayao Chen, Jindong Liu and Juan Wang

Sensors 2025, 25(24), 7608; https://doi.org/10.3390/s25247608 (registering DOI) - 15 Dec 2025

Abstract

With the rapid advancement of speech synthesis and voice conversion technologies, audio deepfake techniques have posed serious threats to information security. Existing detection methods often lack robustness when confronted with environmental noise, signal compression, and ambiguous fake features, making it difficult to effectively [...] Read more.

With the rapid advancement of speech synthesis and voice conversion technologies, audio deepfake techniques have posed serious threats to information security. Existing detection methods often lack robustness when confronted with environmental noise, signal compression, and ambiguous fake features, making it difficult to effectively identify highly concealed fake audio. To address this issue, this paper proposes a Dual-Path Time-Frequency Attention Network (DPTFAN) based on Pythagorean Hesitant Fuzzy Sets (PHFS), which dynamically characterizes the reliability and ambiguity of fake features through uncertainty modeling. It introduces a dual-path attention mechanism in both time and frequency domains to enhance feature representation and discriminative capability. Additionally, a Lightweight Fuzzy Branch Network (LFBN) is designed to achieve explicit enhancement of ambiguous features, improving performance while maintaining computational efficiency. On the ASVspoof 2019 LA dataset, the proposed method achieves an accuracy of 98.94%, and on the FoR (Fake or Real) dataset, it reaches an accuracy of 99.40%, significantly outperforming existing mainstream methods and demonstrating excellent detection performance and robustness. Full article

(This article belongs to the Section Sensor Networks)

► Show Figures

Figure 1

31 pages, 2898 KB

Open AccessArticle

TCN-LSTM-AM Short-Term Photovoltaic Power Forecasting Model Based on Improved Feature Selection and APO

by Ning Ye, Chaoyang Zhi, Yongchao Yu, Sen Lin and Fengxian Liu

Sensors 2025, 25(24), 7607; https://doi.org/10.3390/s25247607 (registering DOI) - 15 Dec 2025

Abstract

The inherent volatility and intermittency of solar power generation pose significant challenges to the stability of power systems. Consequently, high-precision power forecasting is critical for mitigating these impacts and ensuring reliable operation. This paper proposes a framework for photovoltaic (PV) power forecasting that [...] Read more.

The inherent volatility and intermittency of solar power generation pose significant challenges to the stability of power systems. Consequently, high-precision power forecasting is critical for mitigating these impacts and ensuring reliable operation. This paper proposes a framework for photovoltaic (PV) power forecasting that integrates refined feature engineering with deep learning models in a two-stage approach. In the feature engineering stage, a KNN-PCC-SHAP method is constructed. This method is initiated with the KNN algorithm, which is used to identify anomalous samples and perform data interpolation. PCC is then used to screen linearly correlated features. Finally, the SHAP value is used to quantitatively analyze the nonlinear contributions and interaction effects of each feature, thereby forming an optimal feature subset with higher information density. In the modeling stage, a TCN-LSTM-AM combined forecasting model is constructed to collaboratively capture the local details, long-term dependencies, and key timing features of the PV power sequence. The APO algorithm is utilized for the adaptive optimization of the crucial configuration parameters within the model. Experiments based on real PV power plants and public data show that the framework outperforms multiple comparison models in terms of key indicators such as RMSE (2.1098 kW), MAE (1.1073 kW), and R² (0.9775), verifying that the deep integration of refined feature engineering and deep learning models is an effective way to improve the accuracy of PV power prediction. Full article

(This article belongs to the Topic AI and Data-Driven Advancements in Industry 4.0, 2nd Edition)

15 pages, 121628 KB

Open AccessArticle

Deep Guided Exposure Correction with Knowledge Distillation

by Songrong Liu and Tao Zhang

Sensors 2025, 25(24), 7606; https://doi.org/10.3390/s25247606 (registering DOI) - 15 Dec 2025

Abstract

Images captured with unreasonable exposures will greatly reduce visual quality. Exposure problems can be categorized as follows: (i) over-exposure, i.e., bright and losing image regions caused by too-long exposure; (ii) under-exposure, i.e., dark and drowned-in noises caused by too-short exposure. Most prior works [...] Read more.

Images captured with unreasonable exposures will greatly reduce visual quality. Exposure problems can be categorized as follows: (i) over-exposure, i.e., bright and losing image regions caused by too-long exposure; (ii) under-exposure, i.e., dark and drowned-in noises caused by too-short exposure. Most prior works only handle over- or under-exposure problems on sRGB domain and ignore prior knowledge of channel information. In this paper, we propose Deep Guided network for exposure correction on RAW domain with Knowledge Distillation (denoted as DGKD), solving two problems together. Firstly, according to color sensitivity, we employ blue/red channel and green channel as guidance information for over- and under-exposure correction, respectively. Secondly, to handle two varying problems in a unified network, we first train the over- and under-exposure correction networks individually and then distill knowledge into one deep guided network. The experimental results show that the proposed method outperforms the state-of-the-art methods under both quantitative metrics and visual quality. Specifically, the proposed method attained a peak signal-to-noise ratio of 24.653 dB and a structural similarity index of 0.8182 on the collected RAW image exposure correction dataset. Full article

(This article belongs to the Collection Advances in Deep-Learning-Based Sensing, Imaging, and Video Processing)

32 pages, 1628 KB

Open AccessArticle

Anomaly Behavior Detection Based on Deep Learning in an IoT Environment

by Anqi Fu and Jian Li

Sensors 2025, 25(24), 7605; https://doi.org/10.3390/s25247605 (registering DOI) - 15 Dec 2025

Abstract

In the era of the Internet of Things (IoT), video surveillance, as a vital component of smart cities and public security systems, faces the critical challenge of efficiently detecting abnormal behaviors within massive video streams. However, existing weakly supervised video anomaly detection methods [...] Read more.

In the era of the Internet of Things (IoT), video surveillance, as a vital component of smart cities and public security systems, faces the critical challenge of efficiently detecting abnormal behaviors within massive video streams. However, existing weakly supervised video anomaly detection methods are often limited by the scarcity of abnormal samples, the similarity between normal and abnormal segments, and the insufficient modeling of temporal dependencies. To address these challenges, this paper proposes a novel approach that integrates temporal structural attention with contrastive learning. On the one hand, causal masks and temporal decay weights are incorporated into the attention mechanism to explicitly constrain temporal relations and prevent future information leakage; on the other hand, positive/negative offsets and a contrastive learning strategy are employed to enhance the discriminability of abnormal segments in the latent space. Experiments conducted on multiple public video anomaly detection datasets validate the effectiveness of the proposed method, with results showing superior performance over existing mainstream models: the AUC increases to 98.1%, ACC reaches 96.1%, and the F1-score improves to 94.5%. These findings demonstrate that the proposed method can provide more intelligent, efficient, and reliable anomaly detection for IoT-based video surveillance, holding significant implications for public safety and intelligent monitoring. Full article

(This article belongs to the Special Issue IoT Network Security (Second Edition))

18 pages, 4126 KB

Open AccessArticle

Fault Diagnosis of Static Eccentricity in Marine Diesel Generators Using 2D Short-Time Fourier Transform of Three-Phase Currents

by Beom-Jin Joe, Jin-Sung Lee, Sang-Jae Yeo, Yong Jae Cho and Jee-Yeon Jeon

Sensors 2025, 25(24), 7604; https://doi.org/10.3390/s25247604 (registering DOI) - 15 Dec 2025

Abstract

Static eccentricity is an important early-stage fault in marine diesel generators, as small air-gap deviations caused by misalignment or mechanical wear can escalate into bearing damage and rotor–stator contact. To address the challenge of detecting such subtle faults, this study proposes a current [...] Read more.

Static eccentricity is an important early-stage fault in marine diesel generators, as small air-gap deviations caused by misalignment or mechanical wear can escalate into bearing damage and rotor–stator contact. To address the challenge of detecting such subtle faults, this study proposes a current signal analysis method based on the two-dimensional short-time Fourier transform (2D STFT) for early detection of static eccentricity faults in marine diesel generators. Using three-phase currents measured during normal operation and fault data synthesized with a physics-based electromechanical coupling model (1–5% eccentricity), we construct a two-dimensional phase–time representation rather than treating each phase as an independent one-dimensional time series and then apply 2D STFT. This formulation enables the simultaneous capture of inter-phase relationships and spatial patterns in the time–frequency–phase domain. Experiments indicate a distinct energy rise near 1020 Hz as static eccentricity increases. This trend enables the proposed method to distinguish small faults of approximately 5% eccentricity, which remain difficult to detect using conventional 1D STFT. As a result, the approach improves the diagnostic accuracy of non-contact, current-based monitoring for static eccentricity faults. Future work will include validation using real in-service fault data and extensions to other fault modes such as dynamic eccentricity and bearing defects. Full article

(This article belongs to the Special Issue Sensors for Predictive Maintenance of Machines: 2nd Edition)

► Show Figures

Figure 1

26 pages, 8544 KB

Open AccessArticle

Hi-MDTCN: Hierarchical Multi-Scale Dilated Temporal Convolutional Network for Tool Condition Monitoring

by Anying Chai, Zhaobo Fang, Mengjia Lian, Ping Huang, Chenyang Guo, Wanda Yin, Lei Wang, Enqiu He and Siwen Li

Sensors 2025, 25(24), 7603; https://doi.org/10.3390/s25247603 (registering DOI) - 15 Dec 2025

Abstract

Accurate identification of tool wear conditions is of great significance for extending tool life, ensuring processing quality, and improving production efficiency. Current research shows that signals collected by a single sensor have limited dimensions and cannot comprehensively capture the degradation process of tool [...] Read more.

Accurate identification of tool wear conditions is of great significance for extending tool life, ensuring processing quality, and improving production efficiency. Current research shows that signals collected by a single sensor have limited dimensions and cannot comprehensively capture the degradation process of tool wear, while multi-sensor fusion recognition methods cannot effectively handle the complementarity and redundancy between heterogeneous sensor data in feature extraction and fusion. To address these issues, this paper proposes Hi-MDTCN (Hierarchical Multi-scale Dilated Temporal Convolutional Network). In the network, we propose a hierarchical signal analysis framework that processes the signal in segments. When processing intra-segment signals, we design a Multi-channel one-dimensional convolutional network with attention mechanism to capture local wear features at different time scales and fuse them into a unified representation. When processing signal segments, we design a Bi-TCN module to further capture long-term dependencies in wear evolution, mining the overall trend of tool wear over time. Hi-MDTCN adopts a dilated convolution mechanism, which can achieve an extremely large receptive field without building an overly deep network structure, effectively solving problems faced by recurrent neural networks in long sequence modeling such as gradient vanishing, low training efficiency, and poor parallel computing capability, achieving efficient parallel capture of long-range dependencies in time series. Finally, the proposed method is applied to the PHM2010 milling data. Experimental results show that the model’s tool condition recognition accuracy is higher than traditional methods, demonstrating its effectiveness for practical applications. Full article

(This article belongs to the Special Issue Sensing Technologies in Industrial Defect Detection)

► Show Figures

Figure 1

21 pages, 1974 KB

Open AccessArticle

Low-Temperature Stress-Induced Changes in Cucumber Plants—A Near-Infrared Spectroscopy and Aquaphotomics Approach for Investigation

by Daniela Moyankova, Petya Stoykova, Petya Veleva, Nikolai K. Christov, Antoniya Petrova, Krasimir Rusanov and Stefka Atanassova

Sensors 2025, 25(24), 7602; https://doi.org/10.3390/s25247602 (registering DOI) - 15 Dec 2025

Abstract

Low temperatures have a significant impact on the growth, development, and productivity of cucumber plants. The potential of near-infrared spectroscopy and the aquaphotomics approach for investigating chilling stress was studied in Voreas F1 and Gergana cultivars. Changes in the spectral patterns of cucumber [...] Read more.

Low temperatures have a significant impact on the growth, development, and productivity of cucumber plants. The potential of near-infrared spectroscopy and the aquaphotomics approach for investigating chilling stress was studied in Voreas F1 and Gergana cultivars. Changes in the spectral patterns of cucumber plants were compared with physiological and metabolic data. Voreas plants were unable to survive seven days of low-temperature stress due to a drastic increase in electrolyte leakage and a decrease in the net photosynthesis rate, stomatal conductance, and transpiration rate. Gergana plants survived chilling by preserving cell membrane integrity and photosynthesis efficiency. During chilling treatment, the content of most metabolites in both cultivars was reduced compared to the controls, yet it was much more pronounced in Voreas. We observed an increased accumulation of cinnamic acid on the seventh day only in the Gergana cultivar. A MicroNIR spectrometer was used for in vivo spectral measurements of cotyledons and the first two leaves. Differences in absorption spectra were observed among control, stressed, and recovered plants, across different days of stress, and between the studied cultivars. The most significant differences were in the 1300–1600 nm range, much smaller for Gergana than Voreas. Aquagrams of the two cultivars also reveal differences in their responses to low temperatures and changes in water molecular structure in the leaves. The errors of prediction for the days of chilling by using PLS models were from 0.96 to 1.14 days for independent validation, depending on the spectral data of different leaves used. Near-infrared spectroscopy and aquaphotomics can be used as additional tools for early detection of stress and investigation of low-temperature tolerance in cucumber cultivars. Full article

(This article belongs to the Special Issue Spectroscopy and Sensing Technologies for Smart Agriculture)

► Show Figures

Figure 1

20 pages, 4529 KB

Open AccessArticle

Intelligent Recognition of Muffled Blasting Sounds and Lithology Prediction in Coal Mines Based on RDGNet

by Gengxin Li, Hua Ding, Kai Wang, Xiaoqiang Zhang and Jiacheng Sun

Sensors 2025, 25(24), 7601; https://doi.org/10.3390/s25247601 (registering DOI) - 15 Dec 2025

Abstract

In the Yangquan coal mining region, China, muffled blasting sounds commonly occur in mine surrounding rocks resulting from instantaneous energy release following the elastic deformation of overlying brittle rock layers; they are related to fracture development. Although these events rarely cause immediate hazards, [...] Read more.

In the Yangquan coal mining region, China, muffled blasting sounds commonly occur in mine surrounding rocks resulting from instantaneous energy release following the elastic deformation of overlying brittle rock layers; they are related to fracture development. Although these events rarely cause immediate hazards, their acoustic signatures contain critical information about cumulative rock damage. Currently, conventional monitoring of muffled blasting sounds and surrounding rock stability relies on microseismic systems and on-site sampling techniques. However, these methods exhibit low identification efficiency for muffled blasting events, poor real-time performance, and strong subjectivity arising from manual signal interpretation and empirical threshold setting. This article proposes retentive depthwise gated network (RDGNet). By combining retentive network sequence modeling, depthwise separable convolution, and a gated fusion mechanism, RDGNet enables multimodal feature extraction and the fusion of acoustic emission sequences and audio Mel spectrograms, supporting real-time muffled blasting sound recognition and lithology classification. Results confirm model robustness under noisy and multisource mixed-signal conditions (overall accuracy: 92.12%, area under the curve: 0.985, and Macro F1: 0.931). This work provides an efficient approach for intelligent monitoring of coal mine rock stability and can be extended to safety assessments in underground engineering, advancing the mining industry toward preventive management. Full article

(This article belongs to the Section Intelligent Sensors)

► Show Figures

Figure 1

17 pages, 4912 KB

Open AccessArticle

Comparative Study of Distributed Acoustic Sensing Responses in Telecommunication Optical Cables

by Abdulfatah A. G. Abushagur, Mohd Ridzuan Mokhtar, Noor Shafikah Md Rodzi, Khazaimatol Shima Subari, Siti Azlida Ibrahim, Zulkifli Mahmud, Zulfadzli Yusoff, Andre Franzen and Hairul Abdul Rashid

Sensors 2025, 25(24), 7600; https://doi.org/10.3390/s25247600 (registering DOI) - 15 Dec 2025

Abstract

Distributed Acoustic Sensing (DAS) transforms conventional optical fibres into large-scale acoustic sensor arrays. While existing telecommunication cables are increasingly considered for DAS-based monitoring, their performance depends strongly on cable construction and strain transfer efficiency. In this study, the relative DAS signal amplitudes of [...] Read more.

Distributed Acoustic Sensing (DAS) transforms conventional optical fibres into large-scale acoustic sensor arrays. While existing telecommunication cables are increasingly considered for DAS-based monitoring, their performance depends strongly on cable construction and strain transfer efficiency. In this study, the relative DAS signal amplitudes of three commercial telecommunication optical cables were experimentally compared using a benchtop Rayleigh backscattering-based interrogator under controlled laboratory conditions. By maintaining a constant temperature and ensuring no additional strain changes from the outside environment, we guaranteed that only strain-induced variations from acoustic excitations were measured. The results show clear differences in signal amplitude and signal-to-noise ratio (SNR) among the tested cables. The Microcable consistently produced the highest spatial peak amplitude (up to 0.029 a.u.) and SNR (up to 79), while the Duct cable reached 0.00268 a.u. with mean SNR ≈ 32. The Anti-Rodent cable showed low signal amplitude (0.0018 a.u.) but exhibited a high mean SNR (≈111) driven by an exceptional low noise floor in one of the runs. These findings reflect the variations in mechanical coupling between the fibre core and external perturbations and provide practical insights into the suitability of different telecom cable types for DAS applications, supporting informed choices for future deployments. Full article

(This article belongs to the Special Issue Distributed Fibre Optic Sensing Technologies and Applications)

► Show Figures

Figure 1

16 pages, 2677 KB

Open AccessArticle

Outlier-Resistant Initial Alignment of DVL-Aided SINS Using Mahalanobis Distance

by Yidong Shen, Li Luo, Guoqing Wang, Tao Liu, Lin Luo, Jiaxi Guo and Shuangshuang Wang

Sensors 2025, 25(24), 7599; https://doi.org/10.3390/s25247599 (registering DOI) - 15 Dec 2025

Abstract

Due to the influence of the complex underwater environment, the initial alignment method for Doppler velocity log (DVL)-aided strap-down inertial navigation systems (SINS) often suffer from performance degradation, especially when DVL measurements are contaminated by outliers. In this paper, an outlier-resistant Initial Alignment [...] Read more.

Due to the influence of the complex underwater environment, the initial alignment method for Doppler velocity log (DVL)-aided strap-down inertial navigation systems (SINS) often suffer from performance degradation, especially when DVL measurements are contaminated by outliers. In this paper, an outlier-resistant Initial Alignment method with interference suppression for SINS/DVL integrated navigation system is proposed, by which, by constructing an improved Mahalanobis distance anomalous detection criterion, the anomaly of the residual vector composed of observation vectors is judged, and an adaptive weighting factor is introduced into the observation matrix to suppress the abnormal interference in the alignment process. Simulation and experimental results show that, compared with existing initial alignment methods, the proposed method achieves higher alignment accuracy in the presence of outliers, which is more suitable for the SINS/DVL integrated navigation system. Full article

(This article belongs to the Section Navigation and Positioning)

► Show Figures

Figure 1

14 pages, 2795 KB

Open AccessCommunication

Transmission Characteristics of 80 Gbit/s Nyquist-DWDM System in Atmospheric Turbulence

by Silun Du, Qiaochu Yang, Tuo Chen and Tianshu Wang

Sensors 2025, 25(24), 7598; https://doi.org/10.3390/s25247598 (registering DOI) - 15 Dec 2025

Abstract

We experimentally demonstrate an 80 Gbit/s Nyquist-dense wavelength division multiplexed (Nyquist-DWDM) transmission system operating in a simulated atmospheric turbulence channel. The system utilizes eight wavelength-tunable lasers with 100 GHz spacing, modulated by cascaded Mach–Zehnder modulators, to generate phase-locked Nyquist pulse sequences with a [...] Read more.

We experimentally demonstrate an 80 Gbit/s Nyquist-dense wavelength division multiplexed (Nyquist-DWDM) transmission system operating in a simulated atmospheric turbulence channel. The system utilizes eight wavelength-tunable lasers with 100 GHz spacing, modulated by cascaded Mach–Zehnder modulators, to generate phase-locked Nyquist pulse sequences with a 10 GHz repetition rate and a temporal width of 66.7 ps. Each channel is synchronously modulated with a 10 Gbit/s pseudo-random bit sequence (PRBS) and transmitted through controlled weak turbulence conditions generated by a temperature-gradient convection chamber. Experimental measurements reveal that, as the turbulence intensity increases from

C_{n}^{2} = 1.01 \times 10^{- 16}

to

5.71 \times 10^{- 16} m^{- 2 / 3}

, the signal-to-noise ratio (SNR) of the edge channel (C29) and central channel (C33) decreases by approximately 6.5 dB while maintaining stable Nyquist waveform profiles and inter-channel orthogonality. At a forward-error-correction (FEC) threshold of

3.8 \times 10^{- 3}

, the minimum receiver sensitivity is −17.66 dBm, corresponding to power penalties below 5 dB relative to the back-to-back condition. The consistent SNR difference (<2 dB) between adjacent channels confirms uniform power distribution and low inter-channel crosstalk under turbulence. These findings verify that Nyquist pulse shaping substantially mitigates phase distortion and scintillation effects, demonstrating the feasibility of high-capacity DWDM free-space optical (FSO) systems with enhanced spectral efficiency and turbulence resilience. The proposed configuration provides a scalable foundation for future multi-wavelength FSO links and hybrid fiber-wireless optical networks. Full article

(This article belongs to the Special Issue Sensing Technologies and Optical Communication)

► Show Figures

Figure 1

10 pages, 1143 KB

Open AccessArticle

Study on an All-Optic Temperature Sensor Based on a Low-Coherent Optical Interferometry

by Fanni Gu, Yirui Wen, Sergei Krasovskii and Changsen Sun

Sensors 2025, 25(24), 7597; https://doi.org/10.3390/s25247597 (registering DOI) - 15 Dec 2025

Abstract

Optical temperature sensors with intrinsic characteristics of explosion-proof are particularly suitable for the petrochemical industry, etc. However, their applications remain limited by environmental compatibility, etc. Here, we developed an all-optic temperature sensor using an anti-bending single-mode optical fiber in a 3.5 m length [...] Read more.

Optical temperature sensors with intrinsic characteristics of explosion-proof are particularly suitable for the petrochemical industry, etc. However, their applications remain limited by environmental compatibility, etc. Here, we developed an all-optic temperature sensor using an anti-bending single-mode optical fiber in a 3.5 m length and a 0.25 mm outer diameter to match a stainless tube with a 0.4 mm inner diameter. The fiber was threaded into the tube, well bonded with epoxy at both ends of the tube, and configured as one arm of a low-coherent Michelson interferometer. Then, the tube with an embedded sensing fiber was fabricated into a spring, whose final length was about 70 mm with an outside diameter of 13 mm. Changes in temperature alter the lengths of the stainless tube spring in a thermoelastic way, thereby modifying the inner fiber length and producing an optical path difference between the sensing fiber and the packaged reference arm of the interferometer. A temperature calibration was carried out from −25 to 65 °C, and the results demonstrated that the hysteresis of the spring sensor was within ±1.16 °C and the sensitivity was 0.34 °C, which was verified by using a platinum resistance temperature sensor (PT-100). This work provides a reference for further intrinsic optical temperature sensor design. Full article

(This article belongs to the Special Issue Advances and Innovations in Optical Fiber Sensors)

► Show Figures

Figure 1

22 pages, 9457 KB

Open AccessArticle

Enhancing Document Classification Through Multimodal Image-Text Classification: Insights from Fine-Tuned CLIP and Multimodal Deep Fusion

by Hosam Aljuhani, Mohamed Yehia Dahab and Yousef Alsenani

Sensors 2025, 25(24), 7596; https://doi.org/10.3390/s25247596 (registering DOI) - 15 Dec 2025

Abstract

Foundation models excel on general benchmarks but often underperform in clinical settings due to domain shift between internet-scale pretraining data and medical data. Multimodal deep learning, which jointly leverages medical images and clinical text, is promising for diagnosis, yet it remains unclear whether [...] Read more.

Foundation models excel on general benchmarks but often underperform in clinical settings due to domain shift between internet-scale pretraining data and medical data. Multimodal deep learning, which jointly leverages medical images and clinical text, is promising for diagnosis, yet it remains unclear whether domain adaptation is better achieved by fine-tuning large vision–language models or by training lighter, task-specific architectures. We address this question by introducing PairDx, a balanced dataset of 22,665 image–caption pairs spanning six medical document classes, curated to reduce class imbalance and support fair, reproducible comparisons. Using PairDx, we develop and evaluate two approaches: (i) PairDxCLIP, a fine-tuned CLIP (ViT-B/32), and (ii) PairDxFusion, a custom hybrid model that combines ResNet-18 visual features and GloVe text embeddings with attention-based fusion. Both adapted models substantially outperform a zero-shot CLIP baseline (61.18% accuracy) and a specialized model, BiomedCLIP, which serves as an additional baseline and achieves 66.3% accuracy. Our fine-tuned CLIP (PairDxCLIP) attains 93% accuracy and our custom fusion model (PairDxFusion) reaches 94% accuracy on a held-out test set. Notably, PairDxFusion achieves this high accuracy with 17 min, 55 s of training time, nearly four times faster than PairDxCLIP (65 min, 52 s), highlighting a practical efficiency–performance trade-off for clinical deployment. The testing time also outperforms the specialized model—BiomedCLIP (0.387 s/image). Our results demonstrate that carefully constructed domain-specific datasets and lightweight multimodal fusion can close the domain gap while reducing computational cost in healthcare decision support. Full article

(This article belongs to the Special Issue Transforming Healthcare with Smart Sensing and Machine Learning)

► Show Figures

Figure 1

16 pages, 374 KB

Open AccessArticle

Trajectories of Adherence to Study-Prescribed Physical Activity Goals in a mHealth Weight Loss Intervention

by Zhadyra Bizhanova, Lora E. Burke, Maria M. Brooks, Bonny Rockette-Wagner, Jacob K. Kariuki and Susan M. Sereika

Sensors 2025, 25(24), 7595; https://doi.org/10.3390/s25247595 (registering DOI) - 15 Dec 2025

Abstract

Introduction: Engaging in ≥300 min/week of moderate-to-vigorous physical activity (MVPA) is recommended for weight management. This study identified MVPA goal-adherence trajectories and associated predictors and weight outcomes in a 12-month mHealth weight-loss trial. Materials and Methods: This was a secondary analysis [...] Read more.

Introduction: Engaging in ≥300 min/week of moderate-to-vigorous physical activity (MVPA) is recommended for weight management. This study identified MVPA goal-adherence trajectories and associated predictors and weight outcomes in a 12-month mHealth weight-loss trial. Materials and Methods: This was a secondary analysis of valid PA data (≥4 days/week with ≥500 steps/day) from participants (age ≥ 18 years, BMI 27–43 kg/m²) randomized 1:1 to self-monitoring with tailored feedback or self-monitoring only. Both groups received Fitbit trackers. Group-based trajectory modeling identified adherence trajectories and baseline predictors. Analysis of variance was used to estimate associations between trajectory group membership and 12-month weight change. Results: Among 502 participants (79% female, 82% White, mean age of 45.0 ±14.4 years), four MVPA goal-adherence trajectories were identified: lower stable (34.5%), moderate (39.8%), increasing (19.3%), and high (6.4%). A graded association was observed with better adherence trajectories being associated with greater 12-month weight loss (p < 0.0001). Older age, male sex, being unpartnered, and higher first-week MVPA predicted membership in higher adherence trajectory groups (p < 0.05). Conclusions: Higher MVPA goal-adherence was related to greater weight loss. Early MVPA levels predicted long-term adherence, supporting the importance of personalized, technology-supported strategies to promote long-term PA adherence and inform targeted interventions to prevent chronic diseases. Full article

(This article belongs to the Special Issue Sensors for Physiological Monitoring and Digital Health: 2nd Edition)

► Show Figures

Figure 1

20 pages, 6791 KB

Open AccessArticle

A Novel Denoising Method for Mud Continuous-Wave Signals Based on Selective Ensemble Strategy with Particle Swarm Optimization

by Chongjun Huang, Wenbo Cai, Dongxiao Pang, Yan Yang, Xingwang Du, Xi Duan and Yunxu Zhou

Sensors 2025, 25(24), 7594; https://doi.org/10.3390/s25247594 (registering DOI) - 15 Dec 2025

Abstract

During drilling operations, mud continuous-wave signals suffer severe distortion at the surface receiver due to dynamic and complex background noise. Traditional noise cancelation methods face limitations in effectiveness and generalization. To address this, this paper proposes a particle swarm optimization (PSO)-based selective ensemble [...] Read more.

During drilling operations, mud continuous-wave signals suffer severe distortion at the surface receiver due to dynamic and complex background noise. Traditional noise cancelation methods face limitations in effectiveness and generalization. To address this, this paper proposes a particle swarm optimization (PSO)-based selective ensemble filtering strategy. It adaptively selects optimal filtering algorithms matched to the target scenario and assigns their weights, enabling complementary strengths from diverse filters. Preferred filtering methods are applied in parallel, generating initial signals as candidate solutions within the reconstructed ensemble. The PSO algorithm identifies the optimal matching weights for each particle and performs the optimal selective ensemble of these candidates. This process enhances generalization capability and achieves high-precision reconstruction of the continuous-wave signal in complex drilling environments. Experimental results demonstrate that the reconstructed signal using this method exhibits superior quality metrics and greater robustness compared to any single-filter output, validating the proposed approach’s effectiveness. Full article

(This article belongs to the Special Issue Cutting-Edge Developments in Optical Communications, Perception and Computing)

► Show Figures

Figure 1

21 pages, 10300 KB

Open AccessArticle

Cross-Detector Visual Localization with Coplanarity Constraints for Indoor Environments

by Jose-Luis Matez-Bandera, Alberto Jaenal, Clara Gomez, Alejandra C. Hernandez, Javier Monroy, José Araújo and Javier Gonzalez-Jimenez

Sensors 2025, 25(24), 7593; https://doi.org/10.3390/s25247593 (registering DOI) - 15 Dec 2025

Abstract

Most visual localization (VL) methods typically assume that keypoints in the query image are detected with the same algorithm as those stored in the reference map. This poses a serious limitation, as new and better detectors may progressively appear, and we would like [...] Read more.

Most visual localization (VL) methods typically assume that keypoints in the query image are detected with the same algorithm as those stored in the reference map. This poses a serious limitation, as new and better detectors may progressively appear, and we would like to ensure the interoperability and coexistence of cameras with heterogeneous detectors in a single map representation. While rebuilding the map with new detectors might seem a solution, it is often impractical, as original images may be unavailable or restricted due to data privacy constraints. In this paper, we address this challenge with two main contributions. First, we introduce and formalize the problem of cross-detector VL, in which the inherent spatial discrepancies between keypoints from different detectors hinder the process of establishing correct correspondences when relying strictly on the similarity of descriptors for matching. Second, we propose CoplaMatch, the first approach to solve this problem by relaxing strict descriptor similarity and imposing geometric coplanarity constraints. The latter is achieved by leveraging 2D homographies between groups of query and map keypoints. This process involves segmenting planar patches, which is performed offline once for the map, and also in the query image, which adds an extra computational overhead to the VL process, although we demonstrated in our experiments that this does not hinder the online applicability. We extensively validate our proposal through experiments in indoor environments using real-world datasets, demonstrating its effectiveness against two state-of-the-art methods by enabling accurate localization in cross-detector scenarios. Additionally, our work validates the feasibility of cross-detector VL and opens a new direction for the long-term usability of feature-based maps. Full article

(This article belongs to the Special Issue New Solutions in Pattern Recognition and Intelligent Sensors for Mobile Robots)

► Show Figures

Figure 1

22 pages, 7683 KB

Open AccessArticle

Design and Optimization of an Inductive-Stub-Coupled CSRR for Non-Invasive Glucose Sensing

by Zaid A. Abdul Hassain, Malik J. Farhan, Taha A. Elwi and Iulia Andreea Mocanu

Sensors 2025, 25(24), 7592; https://doi.org/10.3390/s25247592 (registering DOI) - 14 Dec 2025

Abstract

This paper presents a high-sensitivity microwave sensor based on a modified Complementary Split Ring Resonator (CSRR) architecture, integrated with inductive stubs, for non-invasive blood glucose monitoring. The proposed sensor is designed to enhance the electric field localization and coupling efficiency by introducing inductive [...] Read more.

This paper presents a high-sensitivity microwave sensor based on a modified Complementary Split Ring Resonator (CSRR) architecture, integrated with inductive stubs, for non-invasive blood glucose monitoring. The proposed sensor is designed to enhance the electric field localization and coupling efficiency by introducing inductive elements that strengthen the perturbation effect caused by glucose concentration changes in the blood. Numerical simulations were conducted using a multilayer finger model to evaluate the sensor’s performance under various glucose levels ranging from 0 to 500 mg/dL. The modified sensor exhibits dual-resonance characteristics and outperforms the conventional CSRR in both frequency and amplitude sensitivity. At an optimized stub gap of 2 mm, which effectively minimizes the capacitive coupling effect of the transmission line and thereby improves the quality factor, the sensor achieves a frequency shift sensitivity of 0.086 MHz/mg/dL and an amplitude sensitivity of 0.02 dB/mg/dL, compared to 0.032 MHz/mg/dL and 0.0116 dB/mg/dL observed in the standard CSRR structure. This confirms a significant enhancement in sensing performance and field confinement due to the optimized inductive loading. These results represent significant enhancements of approximately 168% and 72%, respectively. With its compact design, increased sensitivity, and potential for wearable implementation, the proposed sensor offers a promising platform for continuous, real-time, and non-invasive glucose monitoring in biomedical applications. Full article

(This article belongs to the Section Biosensors)

► Show Figures

Figure 1

15 pages, 4054 KB

Open AccessArticle

An Intuitive Approach to the Optimal Sampling of an Electromagnetic Field

by Marco Donald Migliore

Sensors 2025, 25(24), 7591; https://doi.org/10.3390/s25247591 (registering DOI) - 14 Dec 2025

Abstract

The goal of this article is to provide an intuitive explanation of an electromagnetic field sampling technique, known as optimal field sampling. This explanation provides insight into the physical mechanisms underlying the sampling strategy along observation curves. Full article

(This article belongs to the Special Issue Recent Advances in Antenna Measurement Techniques)

Journal Description

Sensors

Latest Articles

Journal Menu

Journal Browser

Highly Accessed Articles

Latest Books

E-Mail Alert

News

Topics

Conferences

Special Issues

Topical Collections

Further Information

Guidelines

MDPI Initiatives

Follow MDPI