Sensors (Basel). 2024 Dec 21;24(24):8168. doi: 10.3390/s24248168.

ABSTRACT

This systematic review examines EEG-based imagined speech classification, emphasizing directional words essential for development in the brain-computer interface (BCI). This study employed a structured methodology to analyze approaches using public datasets, ensuring systematic evaluation and validation of results. This review highlights the feature extraction techniques that are pivotal to classification performance. These include deep learning, adaptive optimization, and frequency-specific decomposition, which enhance accuracy and robustness. Classification methods were explored by comparing traditional machine learning with deep learning and emphasizing the role of brain lateralization in imagined speech for effective recognition and classification. This study discusses the challenges of generalizability and scalability in imagined speech recognition, focusing on subject-independent approaches and multiclass scalability. Performance benchmarking across various datasets and methodologies revealed varied classification accuracies, reflecting the complexity and variability of EEG signals. This review concludes that challenges remain despite progress, particularly in classifying directional words. Future research directions include improved signal processing techniques, advanced neural network architectures, and more personalized, adaptive BCI systems. This review is critical for future efforts to develop practical communication tools for individuals with speech and motor impairments using EEG-based BCIs.

PMID:39771903 | DOI:10.3390/s24248168

Sensors (Basel). 2024 Dec 21;24(24):8167. doi: 10.3390/s24248167.

ABSTRACT

Freezing of wind turbines causes loss of wind-generated power. Forecasting or prediction of icing on wind turbine blades based on SCADA sensor data allows taking appropriate actions before icing occurs. This paper presents a newly developed deep learning network model named PCTG (Parallel CNN-TCN GRU) for the purpose of high-accuracy and long-term prediction of icing on wind turbine blades. This model combines three networks, the CNN, TCN, and GRU, in order to incorporate both the temporal aspect of SCADA time-series data as well as the dependencies of SCADA variables. The experimentations conducted by using this model and SCADA data from three wind turbines in a wind farm have generated average prediction accuracies of about 97% for prediction horizons of up to 2 days ahead. The developed model is shown to maintain at least 95% prediction accuracy for long prediction horizons of up to 22 days ahead. Furthermore, for another wind farm SCADA dataset, it is shown that the developed PCTG model achieves over 99% icing prediction accuracy 10 days ahead.

PMID:39771901 | DOI:10.3390/s24248167

Sensors (Basel). 2024 Dec 21;24(24):8164. doi: 10.3390/s24248164.

ABSTRACT

Simultaneous localization and mapping (SLAM) techniques can be used to navigate the visually impaired, but the development of robust SLAM solutions for crowded spaces is limited by the lack of realistic datasets. To address this, we introduce InCrowd-VI, a novel visual-inertial dataset specifically designed for human navigation in indoor pedestrian-rich environments. Recorded using Meta Aria Project glasses, it captures realistic scenarios without environmental control. InCrowd-VI features 58 sequences totaling a 5 km trajectory length and 1.5 h of recording time, including RGB, stereo images, and IMU measurements. The dataset captures important challenges such as pedestrian occlusions, varying crowd densities, complex layouts, and lighting changes. Ground-truth trajectories, accurate to approximately 2 cm, are provided in the dataset, originating from the Meta Aria project machine perception SLAM service. In addition, a semi-dense 3D point cloud of scenes is provided for each sequence. The evaluation of state-of-the-art visual odometry (VO) and SLAM algorithms on InCrowd-VI revealed severe performance limitations in these realistic scenarios. Under challenging conditions, systems exceeded the required localization accuracy of 0.5 m and the 1% drift threshold, with classical methods showing drift up to 5-10%. While deep learning-based approaches maintained high pose estimation coverage (>90%), they failed to achieve real-time processing speeds necessary for walking pace navigation. These results demonstrate the need and value of a new dataset to advance SLAM research for visually impaired navigation in complex indoor environments.

PMID:39771900 | DOI:10.3390/s24248164

Sensors (Basel). 2024 Dec 20;24(24):8157. doi: 10.3390/s24248157.

ABSTRACT

Rolling bearings play a crucial role in industrial equipment, and their failure is highly likely to cause a series of serious consequences. Traditional deep learning-based bearing fault diagnosis algorithms rely on large amounts of training data; training and inference processes consume significant computational resources. Thus, developing a lightweight and suitable fault diagnosis algorithm for small samples is particularly crucial. In this paper, we propose a bearing fault diagnosis algorithm based on probabilistic decoupling knowledge distillation and meta-learning (MIX-MPDKD). This algorithm is lightweight and deployable, performing well in small sample scenarios and effectively solving the deployment problem of large networks in resource-constrained environments. Firstly, our model utilizes the Model-Agnostic Meta-Learning algorithm to initialize the parameters of the teacher model and conduct efficient training. Subsequently, by employing the proposed probability-based decoupled knowledge distillation approach, the outstanding performance of the teacher model was imparted to the student model, enabling the student model to converge rapidly in the context of a small sample size. Finally, the Paderborn University dataset was used for meta-training, while the bearing dataset from Case Western Reserve University, along with our laboratory dataset, was used to validate the results. The experimental results demonstrate that the algorithm achieved satisfactory accuracy performance.

PMID:39771892 | DOI:10.3390/s24248157

Sensors (Basel). 2024 Dec 20;24(24):8139. doi: 10.3390/s24248139.

ABSTRACT

We demonstrate high-resolution single-pixel imaging (SPI) in the visible and near-infrared wavelength ranges using an SPI framework that incorporates a novel, dedicated sampling scheme and a reconstruction algorithm optimized for the rapid imaging of highly sparse scenes at the native digital micromirror device (DMD) resolution of 1024 × 768. The reconstruction algorithm consists of two stages. In the first stage, the vector of SPI measurements is multiplied by the generalized inverse of the measurement matrix. In the second stage, we compare two reconstruction approaches: one based on an iterative algorithm and the other on a trained neural network. The neural network outperforms the iterative method when the object resembles the training set, though it lacks the generality of the iterative approach. For images captured at a compression of 0.41 percent, corresponding to a measurement rate of 6.8 Hz with a DMD operating at 22 kHz, the typical reconstruction time on a desktop with a medium-performance GPU is comparable to the image acquisition rate. This allows the proposed SPI method to support high-resolution dynamic SPI in a variety of applications, using a standard SPI architecture with a DMD modulator operating at its native resolution and bandwidth, and enabling the real-time processing of the measured data with no additional delay on a standard desktop PC.

PMID:39771884 | DOI:10.3390/s24248139

Clin Imaging. 2025 Jan 2;119:110400. doi: 10.1016/j.clinimag.2024.110400. Online ahead of print.

ABSTRACT

PURPOSE: To develop an educational, interactive, ultra-high resolution, in vivo magnetic resonance (MR) neurography atlas for direct visualization of the brachial plexus and upper extremity.

METHODS: A total of 16 adult volunteers without known peripheral neuropathy underwent magnetic resonance (MR) neurography of the brachial plexus and upper extremity. To improve vascular suppression, subjects received an intravenous infusion of ferumoxytol. To improve image quality, MR neurography datasets were reconstructed using a deep learning algorithm. The atlas was then developed using a web-based user-interface software, which allowed for labeling of peripheral nerves and muscles, and mapping of muscles to their respective innervation. The user interface was optimized to maximize interactivity and user-friendliness.

RESULTS: Fifteen subjects completed at least one scan with no reported adverse reactions from the ferumoxytol infusions. Adequate vascular suppression was observed in all MR neurography datasets. The images of the brachial plexus and upper extremity included in this atlas allowed for identification and labeling of 177 unique anatomical structures from the neck to the wrist. The atlas was made freely accessible on the internet.

CONCLUSION: A detailed and interactive MR neurography atlas of the brachial plexus and upper extremity was successfully developed to depict small nerves and fascicular detail with unprecedented spatial and contrast resolution. This freely available online resource (https://www.hss.edu/MRNatlas) can be used as an educational tool and clinical reference. The techniques utilized in this project serve as a framework for continued work in expanding the atlas to cover other peripheral nerve territories.

PMID:39765207 | DOI:10.1016/j.clinimag.2024.110400

J Environ Manage. 2025 Jan 6;374:124011. doi: 10.1016/j.jenvman.2024.124011. Online ahead of print.

ABSTRACT

Accurate meteorological observation data is of great importance to human production activities. Meteorological observation systems have been advancing toward automation, intelligence, and informatization. Yet, instrumental malfunctions and unstable sensor node resources could cause significant deviations of data from the actual characteristics it should reflect. To achieve greater data accuracy, early detections of data anomalies, continuous collections and timely transmissions of data are essential. While obvious anomalies can be readily identified, the detection of systematic and gradually emerging anomalies requires further analyses. This study develops an interpretable deep learning method based on an autoencoder (AE), SHapley Additive exPlanations (SHAP) and Bayesian optimization (BO), in order to facilitate prompt and accurate anomaly detections of meteorological observational data. The proposed method can be unfolded into four parts. Firstly, the AE performs anomaly detections based on multidimensional meteorological datasets by marking the data that shows significant reconstruction errors. Secondly, the model evaluates the importance of each meteorological element of the flagged data via SHapley Additive exPlanation (SHAP). Thirdly, a K-sigma based threshold automatic delineation method is employed to obtain reasonable anomaly thresholds that are subject to the data characteristics of different observation sites. Finally, the BO algorithm is adopted to fine-tune difficult hyperparameters, enhancing the model's structure and thus the accuracy of anomaly detection. The practical implication of the proposed model is to inform agricultural production, climate observation, and disaster prevention.

PMID:39765064 | DOI:10.1016/j.jenvman.2024.124011

ACS Sens. 2025 Jan 7. doi: 10.1021/acssensors.4c02616. Online ahead of print.

ABSTRACT

Rapidly detecting hydrogen leaks is critical for the safe large-scale implementation of hydrogen technologies. However, to date, no technically viable sensor solution exists that meets the corresponding response time targets under technically relevant conditions. Here, we demonstrate how a tailored long short-term transformer ensemble model for accelerated sensing (LEMAS) speeds up the response of an optical plasmonic hydrogen sensor by up to a factor of 40 and eliminates its intrinsic pressure dependence in an environment emulating the inert gas encapsulation of large-scale hydrogen installations by accurately predicting its response value to a hydrogen concentration change before it is physically reached by the sensor hardware. Moreover, LEMAS provides a measure for the uncertainty of the predictions that are pivotal for safety-critical sensor applications. Our results advertise the use of deep learning for the acceleration of sensor response, also beyond the realm of plasmonic hydrogen detection.

PMID:39764741 | DOI:10.1021/acssensors.4c02616

Brief Bioinform. 2024 Nov 22;26(1):bbae685. doi: 10.1093/bib/bbae685.

ABSTRACT

Spatial transcriptomics (ST) technologies enable dissecting the tissue architecture in spatial context. To perceive the global contextual information of gene expression patterns in tissue, the spatial dependence of cells must be fully considered by integrating both local and non-local features by means of spatial-context-aware. However, the current ST integration algorithm ignores for ST dropouts, which impedes the spatial-aware of ST features, resulting in challenges in the accuracy and robustness of microenvironmental heterogeneity detecting, spatial domain clustering, and batch-effects correction. Here, we developed an STMGraph, a universal dual-view dynamic deep learning framework that combines dual-remask (MASK-REMASK) with dynamic graph attention model (DGAT) to exploit ST data outperforming pre-existing tools. The dual-remask mechanism masks the embeddings before encoding and decoding, establishing dual-decoding-view to share features mutually. DGAT leverages self-supervision to update graph linkage relationships from two distinct perspectives, thereby generating a comprehensive representation for each node. Systematic benchmarking against 10 state-of-the-art tools revealed that the STMGraph has the optimal performance with high accuracy and robustness on spatial domain clustering for the datasets of diverse ST platforms from multi- to sub-cellular resolutions. Furthermore, STMGraph aggregates ST information cross regions by dual-remask to realize the batch-effects correction implicitly, allowing for spatial domain clustering of ST multi-slices. STMGraph is platform independent and superior in spatial-context-aware to achieve microenvironmental heterogeneity detection, spatial domain clustering, batch-effects correction, and new biological discovery, and is therefore a desirable novel tool for diverse ST studies.

PMID:39764614 | DOI:10.1093/bib/bbae685

Int J Surg. 2025 Jan 7. doi: 10.1097/JS9.0000000000002220. Online ahead of print.

ABSTRACT

Detection of biomarkers of breast cancer incurs additional costs and tissue burden. We propose a deep learning-based algorithm (BBMIL) to predict classical biomarkers, immunotherapy-associated gene signatures, and prognosis-associated subtypes directly from hematoxylin and eosin stained histopathology images. BBMIL showed the best performance among comparative algorithms on the prediction of classical biomarkers, immunotherapy related gene signatures, and subtypes.

PMID:39764584 | DOI:10.1097/JS9.0000000000002220

Hum Reprod Open. 2024 Nov 21;2025(1):hoae070. doi: 10.1093/hropen/hoae070. eCollection 2025.

ABSTRACT

STUDY QUESTION: How accurately can artificial intelligence (AI) models predict sperm retrieval in non-obstructive azoospermia (NOA) patients undergoing micro-testicular sperm extraction (m-TESE) surgery?

SUMMARY ANSWER: AI predictive models hold significant promise in predicting successful sperm retrieval in NOA patients undergoing m-TESE, although limitations regarding variability of study designs, small sample sizes, and a lack of validation studies restrict the overall generalizability of studies in this area.

WHAT IS KNOWN ALREADY: Previous studies have explored various predictors of successful sperm retrieval in m-TESE, including clinical and hormonal factors. However, no consistent predictive model has yet been established.

STUDY DESIGN SIZE DURATION: A comprehensive literature search was conducted following PRISMA-ScR guidelines, covering PubMed and Scopus databases from 2013 to 15 May 2024. Relevant English-language studies were identified using Medical Subject Headings (MeSH) terms. We also used PubMed's 'similar articles' and 'cited by' features for thorough bibliographic screening to ensure comprehensive coverage of relevant literature.

PARTICIPANTS/MATERIALS SETTING METHODS: The review included studies on patients with NOA where AI-based models were used for predicting m-TESE outcomes, by incorporating clinical data, hormonal levels, histopathological evaluations, and genetic parameters. Various machine learning and deep learning techniques, including logistic regression, were employed. The Prediction Model Risk of Bias Assessment Tool (PROBAST) evaluated the bias in the studies, and their quality was assessed using the Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) guidelines, ensuring robust reporting standards and methodological rigor.

MAIN RESULTS AND THE ROLE OF CHANCE: Out of 427 screened articles, 45 met the inclusion criteria, with most using logistic regression and machine learning to predict m-TESE outcomes. AI-based models demonstrated strong potential by integrating clinical, hormonal, and biological factors. However, limitations of the studies included small sample sizes, legal barriers, and challenges in generalizability and validation. While some studies featured larger, multicenter designs, many were constrained by sample size. Most studies had a low risk of bias in participant selection and outcome determination, and two-thirds were rated as low risk for predictor assessment, but the analysis methods varied.

LIMITATIONS REASONS FOR CAUTION: The limitations of this review include the heterogeneity of the included research, potential publication bias and reliance on only two databases (PubMed and Scopus), which may limit the scope of the findings. Additionally, the absence of a meta-analysis prevents quantitative assessment of the consistency of models. Despite this, the review offers valuable insights into AI predictive models for m-TESE in NOA.

WIDER IMPLICATIONS OF THE FINDINGS: The review highlights the potential of advanced AI techniques in predicting successful sperm retrieval for NOA patients undergoing m-TESE. By integrating clinical, hormonal, histopathological, and genetic factors, AI models can enhance decision-making and improve patient outcomes, reducing the number of unsuccessful procedures. However, to further enhance the precision and reliability of AI predictions in reproductive medicine, future studies should address current limitations by incorporating larger sample sizes and conducting prospective validation trials. This continued research and development is crucial for strengthening the applicability of AI models and ensuring broader clinical adoption.

STUDY FUNDING/COMPETING INTERESTS: The authors would like to acknowledge Mashhad University of Medical Sciences, Mashhad, Iran, for financial support (Grant ID: 4020802). The authors declare no competing interests.

REGISTRATION NUMBER: N/A.

PMID:39764557 | PMC:PMC11700607 | DOI:10.1093/hropen/hoae070

EClinicalMedicine. 2024 Nov 11;78:102919. doi: 10.1016/j.eclinm.2024.102919. eCollection 2024 Dec.

ABSTRACT

BACKGROUND: Infant alertness and neurologic changes can reflect life-threatening pathology but are assessed by physical exam, which can be intermittent and subjective. Reliable, continuous methods are needed. We hypothesized that our computer vision method to track movement, pose artificial intelligence (AI), could predict neurologic changes in the neonatal intensive care unit (NICU).

METHODS: We collected video data linked to electroencephalograms (video-EEG) from infants with corrected age less than 1 year at Mount Sinai Hospital in New York City, a level four urban NICU between February 1, 2021 and December 31, 2022. We trained a deep learning pose recognition algorithm on video feeds, labeling 14 anatomic landmarks in 25 frames/infant. We then trained classifiers on anatomic landmarks to predict cerebral dysfunction, diagnosed from EEG readings by an epileptologist, and sedation, defined by the administration of sedative medications.

FINDINGS: We built the largest video-EEG dataset to date (282,301 video minutes, 115 infants) sampled from a diverse patient population. Infant pose was accurately predicted in cross-validation, held-out frames, and held-out infants with respective receiver operating characteristic area under the curves (ROC-AUCs) 0.94, 0.83, 0.89. Median movement increased with age and, after accounting for age, was lower with sedative medications and in infants with cerebral dysfunction (all P < 5 × 10-3, 10,000 permutations). Sedation prediction had high performance on cross-validation, held-out intervals, and held-out infants (ROC-AUCs 0.90, 0.91, 0.87), as did prediction of cerebral dysfunction (ROC-AUCs 0.91, 0.90, 0.76).

INTERPRETATION: We show that pose AI can be applied in an ICU setting and that an EEG diagnosis, cerebral dysfunction, can be predicted from video data alone. Deep learning with pose AI may offer a scalable, minimally invasive method for neuro-telemetry in the NICU.

FUNDING: Friedman Brain Institute Fascitelli Scholar Junior Faculty Grant and Thrasher Research Fund Early Career Award (F.R.). The Clinical and Translational Science Awards (CTSA) grant UL1TR004419 from the National Center for Advancing Translational Sciences. Office of Research Infrastructure of the National Institutes of Health under award number S10OD026880 and S10OD030463.

PMID:39764545 | PMC:PMC11701473 | DOI:10.1016/j.eclinm.2024.102919

Anim Front. 2025 Jan 4;14(6):11-22. doi: 10.1093/af/vfae028. eCollection 2024 Dec.

NO ABSTRACT

PMID:39764526 | PMC:PMC11700597 | DOI:10.1093/af/vfae028

Anim Front. 2025 Jan 4;14(6):68-71. doi: 10.1093/af/vfae026. eCollection 2024 Dec.

NO ABSTRACT

PMID:39764520 | PMC:PMC11700576 | DOI:10.1093/af/vfae026

Natl Sci Rev. 2024 Nov 15;12(1):nwae413. doi: 10.1093/nsr/nwae413. eCollection 2025 Jan.

ABSTRACT

Affordable high-resolution cameras and state-of-the-art computer vision techniques have led to the emergence of various vision-based tactile sensors. However, current vision-based tactile sensors mainly depend on geometric optics or marker tracking for tactile assessments, resulting in limited performance. To solve this dilemma, we introduce optical interference patterns as the visual representation of tactile information for flexible tactile sensors. We propose a novel tactile perception method and its corresponding sensor, combining structural colors from flexible blazed gratings with deep learning. The richer structural colors and finer data processing foster the tactile estimation performance. The proposed sensor has an overall normal force magnitude accuracy of 6 mN, a planar resolution of 79 μm and a contact-depth resolution of 25 μm. This work presents a promising tactile method that combines wave optics, soft materials and machine learning. It performs well in tactile measurement, and can be expanded into multiple sensing fields.

PMID:39764508 | PMC:PMC11702659 | DOI:10.1093/nsr/nwae413

Front Artif Intell. 2024 Dec 20;7:1446693. doi: 10.3389/frai.2024.1446693. eCollection 2024.

ABSTRACT

One of the foremost causes of global healthcare burden is cancer of the gastrointestinal tract. The medical records, lab results, radiographs, endoscopic images, tissue samples, and medical histories of patients with gastrointestinal malignancies provide an enormous amount of medical data. There are encouraging signs that the advent of artificial intelligence could enhance the treatment of gastrointestinal issues with this data. Deep learning algorithms can swiftly and effectively analyze unstructured, high-dimensional data, including texts, images, and waveforms, while advanced machine learning approaches could reveal new insights into disease risk factors and phenotypes. In summary, artificial intelligence has the potential to revolutionize various features of gastrointestinal cancer care, such as early detection, diagnosis, therapy, and prognosis. This paper highlights some of the many potential applications of artificial intelligence in this domain. Additionally, we discuss the present state of the discipline and its potential future developments.

PMID:39764458 | PMC:PMC11701808 | DOI:10.3389/frai.2024.1446693

ArXiv [Preprint]. 2024 Dec 9:arXiv:2412.06717v1.

ABSTRACT

Bankart lesions, or anterior-inferior glenoid labral tears, are diagnostically challenging on standard MRIs due to their subtle imaging features-often necessitating invasive MRI arthrograms (MRAs). This study develops deep learning (DL) models to detect Bankart lesions on both standard MRIs and MRAs, aiming to improve diagnostic accuracy and reduce reliance on MRAs. We curated a dataset of 586 shoulder MRIs (335 standard, 251 MRAs) from 558 patients who underwent arthroscopy. Ground truth labels were derived from intraoperative findings, the gold standard for Bankart lesion diagnosis. Separate DL models for MRAs and standard MRIs were trained using the Swin Transformer architecture, pre-trained on a public knee MRI dataset. Predictions from sagittal, axial, and coronal views were ensembled to optimize performance. The models were evaluated on a 20% hold-out test set (117 MRIs: 46 MRAs, 71 standard MRIs). Bankart lesions were identified in 31.9% of MRAs and 8.6% of standard MRIs. The models achieved AUCs of 0.87 (86% accuracy, 83% sensitivity, 86% specificity) and 0.90 (85% accuracy, 82% sensitivity, 86% specificity) on standard MRIs and MRAs, respectively. These results match or surpass radiologist performance on our dataset and reported literature metrics. Notably, our model's performance on non-invasive standard MRIs matched or surpassed the radiologists interpreting MRAs. This study demonstrates the feasibility of using DL to address the diagnostic challenges posed by subtle pathologies like Bankart lesions. Our models demonstrate potential to improve diagnostic confidence, reduce reliance on invasive imaging, and enhance accessibility to care.

PMID:39764408 | PMC:PMC11703322

ArXiv [Preprint]. 2024 Dec 24:arXiv:2412.18205v1.

ABSTRACT

Biofilms are resistant microbial cell aggregates that pose risks to health and food industries and produce environmental contamination. Accurate and efficient detection and prevention of biofilms are challenging and demand interdisciplinary approaches. This multidisciplinary research reports the application of a deep learning-based artificial intelligence (AI) model for detecting biofilms produced by Pseudomonas aeruginosa with high accuracy. Aptamer DNA templated silver nanocluster (Ag-NC) was used to prevent biofilm formation, which produced images of the planktonic states of the bacteria. Large-volume bright field images of bacterial biofilms were used to design the AI model. In particular, we used U-Net with ResNet encoder enhancement to segment biofilm images for AI analysis. Different degrees of biofilm structures can be efficiently detected using ResNet18 and ResNet34 backbones. The potential applications of this technique are also discussed.

PMID:39764404 | PMC:PMC11703328

Sci Data. 2025 Jan 6;12(1):16. doi: 10.1038/s41597-024-04345-2.

ABSTRACT

Deep learning methods have shown significant potential in tool wear lifecycle analysis. However, there are fewer open source datasets due to the high cost of data collection and equipment time investment. Existing datasets often fail to capture cutting force changes directly. This paper introduces QIT-CEMC, a comprehensive dataset for the full lifecycle of titanium (Ti6Al4V) tool wear. QIT-CEMC utilizes complex circumferential milling paths and employs a rotary dynamometer to directly measure cutting force and torque, alongside multidimensional data from initial wear to severe wear. The dataset consists of 68 different samples with approximately 5 million rows each, includes vibration, sound, cutting force and torque. Detailed wear pictures and measurement values are also provided. It is a valuable resource for time series prediction, anomaly detection, and tool wear studies. We believe QIT-CEMC will be a crucial resource for smart manufacturing research.

PMID:39762327 | DOI:10.1038/s41597-024-04345-2

Sci Rep. 2025 Jan 6;15(1):970. doi: 10.1038/s41598-024-83705-9.

ABSTRACT

Bathymetric mapping of the coastal area is essential for coastal development and management. However, conventional bathymetric measurement in coastal areas is resource-expensive and under many constraints. Various research have been conducted to improve the efficiency or effectiveness of bathymetric estimations. Among them, Satellite-Derived Bathymetry (SDB) shows the greatest promise in providing a cost-effective and efficient solution due to the spatial and temporal resolution offered by satellite imagery. However, the majority of the SDB models are designed for regional bathymetry, which requires prior knowledge of the tested region. This strongly constrains their application to other regions. In this work, we present TransBathy, a deep-learning-based satellite-derived bathymetric model, to solve the coastal bathymetric mapping for different unknown challenging terrains. This model is purposefully crafted to simultaneously assimilate deep and spatial features by employing an attention mechanism. In addition, we collected a large-scale bathymetric dataset covering different shallow coastal regions across the world, including Honolulu Island, Abu Dhabi, Puerto Rico, etc. We trained the model using the collected dataset in an end-to-end manner. We validated the robustness and effectiveness of our model by conducting extensive experiments, including pre-seen and un-seen regions bathymetric estimations. When testing on pre-seen coastal regions in different locations of the world, our model achieves a good performance with an RMSE [Formula: see text] m and R2 [Formula: see text] in the depth down to [Formula: see text] m. When testing in challenging unseen coastal regions with different bottom types, our model obtains RMSE [Formula: see text] m and R2 [Formula: see text] in the steep slope region with depth down to [Formula: see text] m and obtains RMSE [Formula: see text] m and R2 [Formula: see text] in the rugged region with depth down to [Formula: see text] m. Our model even surpasses the baseline SDB method that is pre-trained in these regions by substantially reducing the RMSE by 0.978m and improving the R2 by 0.187 in the steep slope region. The dataset, code, and trained weights of the model are publicly available on GitHub.

PMID:39762308 | DOI:10.1038/s41598-024-83705-9