Noninvasive phonocardiogram (PCG) offers significant potential for the detection of heart failure with preserved ejection fraction (HFpEF), a complex and heterogeneous clinical syndrome. However, the limitations of clinical data have hindered extensive research in this area. While deep learning has shown superior performance over traditional machine learning, its reliance on large-scale annotated datasets and vulnerability to catastrophic forgetting during incremental learning remain notable challenges. To address these challenges, this study proposes a novel continual learning (CL) framework, PCG-CL, which incorporates training strategy and is built upon a lightweight deep network (0.61 M parameters) augmented with a single-head Transformer. This design enables the coordinated facilitation of global and local features, improving the recognition of heart sound (HS) characteristics, while enabling adaptive knowledge retention as clinical data expands incrementally. Additionally, inspired by curriculum learning, this article categorizes data based on varying levels of difficulty and trains them sequentially to facilitate knowledge transfer between tasks. Evaluations on clinical PCG data show that PCG-CL achieves an accuracy of 94.52%, outperforming MobileNetV2 by 8.19%, while using only 27% of the parameters of MobileNetV2. These findings help address the challenges of the clinical diagnostic gray zone for HFpEF, providing a promising solution for clinical practice.
{"title":"PCG-CL: A Lightweight Transformer-Enhanced Continual Learning Approach for HFpEF Detection","authors":"Lianhuan Wei;Meiling Qiu;Xu Liu;Yineng Zheng;Xingming Guo","doi":"10.1109/TIM.2025.3644561","DOIUrl":"https://doi.org/10.1109/TIM.2025.3644561","url":null,"abstract":"Noninvasive phonocardiogram (PCG) offers significant potential for the detection of heart failure with preserved ejection fraction (HFpEF), a complex and heterogeneous clinical syndrome. However, the limitations of clinical data have hindered extensive research in this area. While deep learning has shown superior performance over traditional machine learning, its reliance on large-scale annotated datasets and vulnerability to catastrophic forgetting during incremental learning remain notable challenges. To address these challenges, this study proposes a novel continual learning (CL) framework, PCG-CL, which incorporates training strategy and is built upon a lightweight deep network (0.61 M parameters) augmented with a single-head Transformer. This design enables the coordinated facilitation of global and local features, improving the recognition of heart sound (HS) characteristics, while enabling adaptive knowledge retention as clinical data expands incrementally. Additionally, inspired by curriculum learning, this article categorizes data based on varying levels of difficulty and trains them sequentially to facilitate knowledge transfer between tasks. Evaluations on clinical PCG data show that PCG-CL achieves an accuracy of 94.52%, outperforming MobileNetV2 by 8.19%, while using only 27% of the parameters of MobileNetV2. These findings help address the challenges of the clinical diagnostic gray zone for HFpEF, providing a promising solution for clinical practice.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-12"},"PeriodicalIF":5.9,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904283","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
This article proposes a novel 3-D Moiré-based visual marker with an explicit geometric model. The proposed marker is designed for out-of-plane rotation measurement, which comprises two periodic masks etched on the opposite sides of a glass wafer. The masks project a Moiré pattern on the image plane of the observing camera, and this Moiré pattern is dramatically sensitive to out-of-plane rotation. The key contribution is that we first explicitly derive, for the first time, the geometric relationship between the rotation angle and the Moiré pattern to build a straightforward measurement model, which reveals that the angle is simply determined by the Moiré phase at the image principal point. The accuracy is independent of the observation distance and camera intrinsic parameters. Estimation and calibration algorithms are given. Experiment demonstrates the superiority, which shows an accuracy that is up to 7 times higher than traditional visual markers and at least 2 times higher than state-of-the-art 3-D Moiré-based markers.
{"title":"Visual-Based Out-of-Plane Rotation Measurement Using 3-D Moiré-Based Marker","authors":"Mingzhu Zhu;Maocan Wu;Mingxuan Wei;Bingwei He;Jiantao Liu;Junzhi Yu","doi":"10.1109/TIM.2025.3644552","DOIUrl":"https://doi.org/10.1109/TIM.2025.3644552","url":null,"abstract":"This article proposes a novel 3-D Moiré-based visual marker with an explicit geometric model. The proposed marker is designed for out-of-plane rotation measurement, which comprises two periodic masks etched on the opposite sides of a glass wafer. The masks project a Moiré pattern on the image plane of the observing camera, and this Moiré pattern is dramatically sensitive to out-of-plane rotation. The key contribution is that we first explicitly derive, for the first time, the geometric relationship between the rotation angle and the Moiré pattern to build a straightforward measurement model, which reveals that the angle is simply determined by the Moiré phase at the image principal point. The accuracy is independent of the observation distance and camera intrinsic parameters. Estimation and calibration algorithms are given. Experiment demonstrates the superiority, which shows an accuracy that is up to 7 times higher than traditional visual markers and at least 2 times higher than state-of-the-art 3-D Moiré-based markers.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-9"},"PeriodicalIF":5.9,"publicationDate":"2025-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904299","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The ultrashort baseline (USBL) system plays an important role in underwater vehicle positioning. The angular misalignment (AM) between the acoustic array and attitude sensor is the main source of error in the USBL system. Existing AM calibration methods are primarily designed for surface platform (SP) and rely on high-precision platform position information provided by satellite positioning devices. However, this requirement is especially strict for platforms performing covert underwater missions. To address this limitation, this article proposes a novel AM calibration method for underwater platform (UP) based on dual seabed datums. The proposed method eliminates the need for platform position parameters by differentially positioning the dual seabed datums using the USBL system. Consequently, it can be conducted entirely underwater, ensuring the covert operation of the UP. To validate its performance, precision analysis, simulations, and field trials were conducted. The results demonstrate that the proposed method achieves calibration precision comparable to that of the existing method for SP, without relying on platform position information.
{"title":"An Angular Misalignment Calibration Method for Underwater Platform Based on Dual Seabed Datums","authors":"Heng Cai;Dajun Sun;Cuie Zheng;Qianzhou Bai;Jucheng Zhang","doi":"10.1109/TIM.2025.3643070","DOIUrl":"https://doi.org/10.1109/TIM.2025.3643070","url":null,"abstract":"The ultrashort baseline (USBL) system plays an important role in underwater vehicle positioning. The angular misalignment (AM) between the acoustic array and attitude sensor is the main source of error in the USBL system. Existing AM calibration methods are primarily designed for surface platform (SP) and rely on high-precision platform position information provided by satellite positioning devices. However, this requirement is especially strict for platforms performing covert underwater missions. To address this limitation, this article proposes a novel AM calibration method for underwater platform (UP) based on dual seabed datums. The proposed method eliminates the need for platform position parameters by differentially positioning the dual seabed datums using the USBL system. Consequently, it can be conducted entirely underwater, ensuring the covert operation of the UP. To validate its performance, precision analysis, simulations, and field trials were conducted. The results demonstrate that the proposed method achieves calibration precision comparable to that of the existing method for SP, without relying on platform position information.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-12"},"PeriodicalIF":5.9,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886563","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1109/TIM.2025.3643068
Yuda Chen;Jingru Wang;Xuan Cao;Genyuan Xing;Yan Liu
Developing electromagnetic vibroseis is an important approach to achieve portable seismic sources for shallow seismic exploration. As an electromechanical device, the performance of the portable electromagnetic vibroseis is determined by its internal electromechanical coupling. Obtaining the force factor distribution curve $text {BL}(x)$ is the basis of studying the electromechanical coupling. Existing methods, such as finite element method and quasi-static method, suffer from reliance on prior information, operational complexity, or limited accuracy. Aiming at these disadvantages in the process of obtaining the $text {BL}(x)$ , a novel indirect measurement method is proposed. The proposed method relies solely on acquisition data, including acceleration, displacement, and induced electromotive force (EMF) of the internal driving coils moving in the air gap magnetic field. This approach enables the acquisition of the full-stroke $text {BL}(x)$ in a single, rapid measurement, offering advantages in operational convenience and measurement accuracy. To suppress the noise in motion signals and induced EMF signal, Kalman and Savitzky–Golay (S–G) filtering methods were, respectively, used. The motion velocity was obtained by integrating filtered acceleration signal, followed by a two-stage calibration process. The $text {BL}(t)$ was obtained by solving the motion velocity and induced EMF. The filtered displacement signal was used to map the $text {BL}(t)$ to full-stroke $text {BL}(x)$ . Experiments were carried out based on a portable electromagnetic vibroseis platform. The experimental results show that the average root-mean-square error (RMSE) of the 12 repeated experiments is 0.3139. Compared to other measurement methods, the proposed method achieves both accuracy and operational convenience in portable electromagnetic vibroseis.
{"title":"A Novel Measurement Method of Electromechanical Coupling in Portable Electromagnetic Vibroseis","authors":"Yuda Chen;Jingru Wang;Xuan Cao;Genyuan Xing;Yan Liu","doi":"10.1109/TIM.2025.3643068","DOIUrl":"https://doi.org/10.1109/TIM.2025.3643068","url":null,"abstract":"Developing electromagnetic vibroseis is an important approach to achieve portable seismic sources for shallow seismic exploration. As an electromechanical device, the performance of the portable electromagnetic vibroseis is determined by its internal electromechanical coupling. Obtaining the force factor distribution curve <inline-formula> <tex-math>$text {BL}(x)$ </tex-math></inline-formula> is the basis of studying the electromechanical coupling. Existing methods, such as finite element method and quasi-static method, suffer from reliance on prior information, operational complexity, or limited accuracy. Aiming at these disadvantages in the process of obtaining the <inline-formula> <tex-math>$text {BL}(x)$ </tex-math></inline-formula>, a novel indirect measurement method is proposed. The proposed method relies solely on acquisition data, including acceleration, displacement, and induced electromotive force (EMF) of the internal driving coils moving in the air gap magnetic field. This approach enables the acquisition of the full-stroke <inline-formula> <tex-math>$text {BL}(x)$ </tex-math></inline-formula> in a single, rapid measurement, offering advantages in operational convenience and measurement accuracy. To suppress the noise in motion signals and induced EMF signal, Kalman and Savitzky–Golay (S–G) filtering methods were, respectively, used. The motion velocity was obtained by integrating filtered acceleration signal, followed by a two-stage calibration process. The <inline-formula> <tex-math>$text {BL}(t)$ </tex-math></inline-formula> was obtained by solving the motion velocity and induced EMF. The filtered displacement signal was used to map the <inline-formula> <tex-math>$text {BL}(t)$ </tex-math></inline-formula> to full-stroke <inline-formula> <tex-math>$text {BL}(x)$ </tex-math></inline-formula>. Experiments were carried out based on a portable electromagnetic vibroseis platform. The experimental results show that the average root-mean-square error (RMSE) of the 12 repeated experiments is 0.3139. Compared to other measurement methods, the proposed method achieves both accuracy and operational convenience in portable electromagnetic vibroseis.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-11"},"PeriodicalIF":5.9,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1109/TIM.2025.3643033
Mainak Chakraborty;Sahil Anchal;Chandan;Bodhibrata Mukhopadhyay;Subrat Kar
This article introduces a large-scale, nonintrusive person identification (PrID) framework using footstep-induced structural vibration signals. The increasing adoption of structural vibration analysis for PrID comes from its inherent nonintrusiveness and privacy preserving characteristics. However, the existing methodologies are often constrained by the scarcity of extensive datasets, both in terms of the number of subjects and the temporal length of individual recordings, and frequently rely on supervised learning paradigms coupled with manual feature engineering. Consequently, the generalization capabilities of these approaches to broader populations are typically limited. To address these limitations, we have curated a comprehensive dataset of structural vibration signals acquired from 100 individuals. In addition, we have developed an unsupervised event detection method using the features based on time, frequency, and wavelet analysis. Furthermore, we have developed DeepStep, a residual attention-based framework specifically designed for efficient feature extraction and classification of structural vibration signals. Experimental evaluation on our curated dataset demonstrates that the proposed approach achieves a Rank-1 accuracy of approximately 92% and a Rank-5 accuracy of approximately 96%.
{"title":"DeepStep: A Deep Learning-Based Indoor Person Identification Framework Using Footstep-Induced Structural Vibration Signals","authors":"Mainak Chakraborty;Sahil Anchal;Chandan;Bodhibrata Mukhopadhyay;Subrat Kar","doi":"10.1109/TIM.2025.3643033","DOIUrl":"https://doi.org/10.1109/TIM.2025.3643033","url":null,"abstract":"This article introduces a large-scale, nonintrusive person identification (PrID) framework using footstep-induced structural vibration signals. The increasing adoption of structural vibration analysis for PrID comes from its inherent nonintrusiveness and privacy preserving characteristics. However, the existing methodologies are often constrained by the scarcity of extensive datasets, both in terms of the number of subjects and the temporal length of individual recordings, and frequently rely on supervised learning paradigms coupled with manual feature engineering. Consequently, the generalization capabilities of these approaches to broader populations are typically limited. To address these limitations, we have curated a comprehensive dataset of structural vibration signals acquired from 100 individuals. In addition, we have developed an unsupervised event detection method using the features based on time, frequency, and wavelet analysis. Furthermore, we have developed DeepStep, a residual attention-based framework specifically designed for efficient feature extraction and classification of structural vibration signals. Experimental evaluation on our curated dataset demonstrates that the proposed approach achieves a Rank-1 accuracy of approximately 92% and a Rank-5 accuracy of approximately 96%.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-11"},"PeriodicalIF":5.9,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145879958","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-11DOI: 10.1109/TIM.2025.3643086
Yuhan Huang;Wentao Huang;Zhengjie Liu;Yu Zhang
Dual-rotor systems are critical components in aeroengines, where failures can lead to severe operational disruptions and significant economic losses. However, the high reliability of these systems results in limited fault data, creating a typical small-sample problem. Due to the uniqueness of dual-rotor systems, small-sample fault diagnosis faces two distinct challenges: 1) significant attenuation of fault characteristics during signal transmission and 2) complex dynamic modeling due to asynchronous vibration coupling and structural nonlinearity. To address these challenges, this article presents the first simulation-to-real transformation framework for dual-rotor systems. Specifically, an asymmetric Gaussian chirplet model (AGCM) is developed to preprocess data and enhance fault characteristics in the time–frequency domain, addressing the feature attenuation issue. For complex dynamic modeling, we propose a two-step approach: first, employing a Hertzian contact theory-based simulation model to generate labeled fault data. Nevertheless, due to the inherent complexity of real systems, a significant distribution gap exists between simulation and real data. To bridge this gap, we introduce an innovative adaptive multiscale style transfer network (AMSTN) to embed real-world style characteristics while preserving critical fault features. Experimental results demonstrate the framework’s superior performance under small-sample conditions.
{"title":"A Simulation-to-Real Transformation for Small-Sample Fault Diagnosis in Aeroengine Dual-Rotor Systems","authors":"Yuhan Huang;Wentao Huang;Zhengjie Liu;Yu Zhang","doi":"10.1109/TIM.2025.3643086","DOIUrl":"https://doi.org/10.1109/TIM.2025.3643086","url":null,"abstract":"Dual-rotor systems are critical components in aeroengines, where failures can lead to severe operational disruptions and significant economic losses. However, the high reliability of these systems results in limited fault data, creating a typical small-sample problem. Due to the uniqueness of dual-rotor systems, small-sample fault diagnosis faces two distinct challenges: 1) significant attenuation of fault characteristics during signal transmission and 2) complex dynamic modeling due to asynchronous vibration coupling and structural nonlinearity. To address these challenges, this article presents the first simulation-to-real transformation framework for dual-rotor systems. Specifically, an asymmetric Gaussian chirplet model (AGCM) is developed to preprocess data and enhance fault characteristics in the time–frequency domain, addressing the feature attenuation issue. For complex dynamic modeling, we propose a two-step approach: first, employing a Hertzian contact theory-based simulation model to generate labeled fault data. Nevertheless, due to the inherent complexity of real systems, a significant distribution gap exists between simulation and real data. To bridge this gap, we introduce an innovative adaptive multiscale style transfer network (AMSTN) to embed real-world style characteristics while preserving critical fault features. Experimental results demonstrate the framework’s superior performance under small-sample conditions.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"75 ","pages":"1-23"},"PeriodicalIF":5.9,"publicationDate":"2025-12-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11298273","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145904279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-12-03DOI: 10.1109/TIM.2025.3632055
Sujaya Das Gupta;Sumit Ghosh;Stanley Johnson;Sankar Majhi;Sankalpa Banerjee;Subhadeep De
Typographical unit errors were introduced in the published version that changed several occurrences of “mHz” (millihertz) to “MHz” (megahertz). The following corrections are made to the published article. 1)Abstract: “greatly enhanced to 10 MHz with ML” should read: “greatly enhanced to 10 mHz with ML.”2)Abstract: “phase drifts of 1.3 MHz” should read: “phase drifts of 1.3 mHz.” 3)Section II-C, Frequency Resolution: “The lowest frequency-tuning resolution of 100 MHz” should read: “The lowest frequency-tuning resolution of 100 mHz.”4)Section II-C, Frequency Resolution: “10-MHz frequency-tuning resolution” should read: “10-mHz frequency-tuning resolution.”5)Section II-C, Frequency Drifts: “drift of 1.3 MHz over 24 h” should read: “drift of 1.3 mHz over 24 h.”6)Section IV, Conclusion: “with its 10-MHz frequency” should read: “with its 10-mHz frequency.”
{"title":"Erratum to “Scalable Design of an Atomic Clock Stabilized and ML-Optimized RF Synthesizer”","authors":"Sujaya Das Gupta;Sumit Ghosh;Stanley Johnson;Sankar Majhi;Sankalpa Banerjee;Subhadeep De","doi":"10.1109/TIM.2025.3632055","DOIUrl":"https://doi.org/10.1109/TIM.2025.3632055","url":null,"abstract":"Typographical unit errors were introduced in the published version that changed several occurrences of “mHz” (millihertz) to “MHz” (megahertz). The following corrections are made to the published article. 1)Abstract: “greatly enhanced to 10 MHz with ML” should read: “greatly enhanced to 10 mHz with ML.”2)Abstract: “phase drifts of 1.3 MHz” should read: “phase drifts of 1.3 mHz.” 3)Section II-C, Frequency Resolution: “The lowest frequency-tuning resolution of 100 MHz” should read: “The lowest frequency-tuning resolution of 100 mHz.”4)Section II-C, Frequency Resolution: “10-MHz frequency-tuning resolution” should read: “10-mHz frequency-tuning resolution.”5)Section II-C, Frequency Drifts: “drift of 1.3 MHz over 24 h” should read: “drift of 1.3 mHz over 24 h.”6)Section IV, Conclusion: “with its 10-MHz frequency” should read: “with its 10-mHz frequency.”","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-1"},"PeriodicalIF":5.9,"publicationDate":"2025-12-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11275705","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145674689","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Conventional phased-array ultrasound can detect defects in nuts; however, accurately reconstructing their irregular shapes and precise spatial structures remains challenging. To faithfully recover the structural distribution of nut cross sections, a dedicated ultrasound tomographic imaging system and a corresponding reconstruction method were developed to generate spatially resolved images of nuts with irregular surfaces. The imaging process includes three steps. First, as the nut rotates on the experimental platform, the ultrasonic array elements transmit and receive signals to form a signal matrix. Second, the collected sparse data are interpolated using the proposed adaptive interpolation algorithm and then reconstructed into an image through filtered back-projection. Finally, the reconstructed image is processed with a diffusion modelbased super resolution (SR) algorithm to produce a high-resolution, large-scale tomographic image. Employing a 5 MHz, 64-element linear array with water as the coupling medium for signal acquisition, the proposed imaging algorithm achieves optimal structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR) values of 0.961 and 29.264 after adaptive interpolation under noise-free conditions. Following SR processing, it attains superior no-reference quality scores with natural image quality evaluator (NIQE), CLIP-based image quality assessment (CLIPIQA), and blind/reference-less image spatial quality evaluator (BRISQUE) scores of 2.4933, 0.6655, and 34.5602, outperforming conventional SR methods across these metrics. These results demonstrate superior performance in image quality. Physical experiments further indicate that the system can produce high-precision tomographic images of nuts with minimal signal sampling, transmission, and storage, highlighting its practical application potential.
{"title":"Ultrasonic Tomography System for Nut Defect Detection Using Linear Arrays","authors":"Shiyuan He;Jianhong Yang;Chuanjiang Hu;Xuejin Zhou;Huaiying Fang","doi":"10.1109/TIM.2025.3638929","DOIUrl":"https://doi.org/10.1109/TIM.2025.3638929","url":null,"abstract":"Conventional phased-array ultrasound can detect defects in nuts; however, accurately reconstructing their irregular shapes and precise spatial structures remains challenging. To faithfully recover the structural distribution of nut cross sections, a dedicated ultrasound tomographic imaging system and a corresponding reconstruction method were developed to generate spatially resolved images of nuts with irregular surfaces. The imaging process includes three steps. First, as the nut rotates on the experimental platform, the ultrasonic array elements transmit and receive signals to form a signal matrix. Second, the collected sparse data are interpolated using the proposed adaptive interpolation algorithm and then reconstructed into an image through filtered back-projection. Finally, the reconstructed image is processed with a diffusion modelbased super resolution (SR) algorithm to produce a high-resolution, large-scale tomographic image. Employing a 5 MHz, 64-element linear array with water as the coupling medium for signal acquisition, the proposed imaging algorithm achieves optimal structural similarity index (SSIM) and peak signal-to-noise ratio (PSNR) values of 0.961 and 29.264 after adaptive interpolation under noise-free conditions. Following SR processing, it attains superior no-reference quality scores with natural image quality evaluator (NIQE), CLIP-based image quality assessment (CLIPIQA), and blind/reference-less image spatial quality evaluator (BRISQUE) scores of 2.4933, 0.6655, and 34.5602, outperforming conventional SR methods across these metrics. These results demonstrate superior performance in image quality. Physical experiments further indicate that the system can produce high-precision tomographic images of nuts with minimal signal sampling, transmission, and storage, highlighting its practical application potential.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-12"},"PeriodicalIF":5.9,"publicationDate":"2025-12-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145879956","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-27DOI: 10.1109/TIM.2025.3637988
Jiayi Zhou;Xiaoli Wang;Weihua Gui;Chunhua Yang;Stephen George Pooley
This study presents a novel soft sensor modeling algorithm for industrial processes, known as the hierarchical attention-based quadruple S (HAQS) model, specifically designed to uncover nonlinear dynamic features within semi-supervised process data. It integrates spatial and process temporal attention with an LSTM layer during encoding, enabling the learning of spatio-process-temporal features. The model utilizes an unsupervised decoder to reconstruct the input data sequence, facilitating the understanding of the intrinsic features of the input data. During the supervised decoding phase, the predicted value of the key variable is fed into the subsequent LSTM cell. This enables the model to learn effectively from a limited amount of key variable data. The HAQS model displayed superior performance in prediction accuracy and stability, outperforming other models like the semi-supervised dynamic feature extracting (SSDFE) network in a practical case study involving a mineral processing grinding-classification circuit. The HAQS model has demonstrated substantial promise for real-world application. Its ability to extract features from complex industrial datasets, along with its semi-supervised learning capabilities, makes it a powerful tool for the optimization of industrial processes.
{"title":"Hierarchical Attention-Based Semi-Supervised Sequence-to-Sequence Soft Sensor Model for Complex Industrial Processes","authors":"Jiayi Zhou;Xiaoli Wang;Weihua Gui;Chunhua Yang;Stephen George Pooley","doi":"10.1109/TIM.2025.3637988","DOIUrl":"https://doi.org/10.1109/TIM.2025.3637988","url":null,"abstract":"This study presents a novel soft sensor modeling algorithm for industrial processes, known as the hierarchical attention-based quadruple S (HAQS) model, specifically designed to uncover nonlinear dynamic features within semi-supervised process data. It integrates spatial and process temporal attention with an LSTM layer during encoding, enabling the learning of spatio-process-temporal features. The model utilizes an unsupervised decoder to reconstruct the input data sequence, facilitating the understanding of the intrinsic features of the input data. During the supervised decoding phase, the predicted value of the key variable is fed into the subsequent LSTM cell. This enables the model to learn effectively from a limited amount of key variable data. The HAQS model displayed superior performance in prediction accuracy and stability, outperforming other models like the semi-supervised dynamic feature extracting (SSDFE) network in a practical case study involving a mineral processing grinding-classification circuit. The HAQS model has demonstrated substantial promise for real-world application. Its ability to extract features from complex industrial datasets, along with its semi-supervised learning capabilities, makes it a powerful tool for the optimization of industrial processes.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-14"},"PeriodicalIF":5.9,"publicationDate":"2025-11-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886535","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2025-11-21DOI: 10.1109/TIM.2025.3635810
Zhifeng Zhang;Tian Zhou;Weidong Du;Qijia Guo
Errors in acoustic arrays can degrade detection performance by compromising the accuracy of direction-of-arrival (DOA) estimation and reducing processing gain. Most conventional array calibration methods are based on far-field conditions, which are challenging to implement in confined spaces. In order to address this problem, we propose a self-calibration method for compensating gain-phase and position errors in linear acoustic arrays under near-field conditions. In this method, the DOA of the signal is used to estimate the array element positions, which are then fit to the actual array, while the relative spatial positions of the source and the array are employed to refine the DOA estimation. This alternating iterative procedure enables the accurate estimation of both the DOA and array element errors. Simulation results confirm the effectiveness of the proposed method. Tank test results demonstrate that the accuracy of DOA estimation after array calibration is improved by an average of 0.2°, and the peak sidelobe ratio (PSLR) of the beam pattern is reduced by an average of 2.03 dB.
{"title":"A Near-Field Gain-Phase and Position Errors Calibration Method for Acoustic Arrays","authors":"Zhifeng Zhang;Tian Zhou;Weidong Du;Qijia Guo","doi":"10.1109/TIM.2025.3635810","DOIUrl":"https://doi.org/10.1109/TIM.2025.3635810","url":null,"abstract":"Errors in acoustic arrays can degrade detection performance by compromising the accuracy of direction-of-arrival (DOA) estimation and reducing processing gain. Most conventional array calibration methods are based on far-field conditions, which are challenging to implement in confined spaces. In order to address this problem, we propose a self-calibration method for compensating gain-phase and position errors in linear acoustic arrays under near-field conditions. In this method, the DOA of the signal is used to estimate the array element positions, which are then fit to the actual array, while the relative spatial positions of the source and the array are employed to refine the DOA estimation. This alternating iterative procedure enables the accurate estimation of both the DOA and array element errors. Simulation results confirm the effectiveness of the proposed method. Tank test results demonstrate that the accuracy of DOA estimation after array calibration is improved by an average of 0.2°, and the peak sidelobe ratio (PSLR) of the beam pattern is reduced by an average of 2.03 dB.","PeriodicalId":13341,"journal":{"name":"IEEE Transactions on Instrumentation and Measurement","volume":"74 ","pages":"1-11"},"PeriodicalIF":5.9,"publicationDate":"2025-11-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145886566","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：