Deep Learning Approach Research Articles

Reduced Complexity Deep Learning Approach for Bluetooth Ranging in Multipath Environments

Reduced Complexity Deep Learning Approach for Bluetooth Ranging in Multipath Environments

Hybrid Deep Learning Approaches for Human Activity Recognition and Postural Transitions Using Mobile Device Sensors

Hybrid Deep Learning Approaches for Human Activity Recognition and Postural Transitions Using Mobile Device Sensors

Optimizing Distributed Energy Resources in Microgrid SCUC through Seq2Seq Scheduling Algorithms

In order to optimize Distributed Energy Resources (DERs) inside a microgrid's Security-Constrained Unit Commitment framework, this study investigates the use of Seq2Seq (Sequence-to-Sequence) scheduling methods (SCUC). There is a growing consensus that microgrids are an important part of the future of the electric grid because of the advantages they provide in terms of reliability, renewable energy integration, and overall efficiency. In the context of complicated SCUC issues, efficient scheduling and optimization of DERs are essential for realizing their full potential. Seq2Seq models, a kind of deep learning approach, have shown impressive performance in several applications requiring sequence prediction. Uncertainty in renewable energy production, energy demand forecasts, and security limitations are all addressed in this work as Seq2Seq algorithms are applied to microgrid SCUC. Significant gains in economic efficiency, ecological viability, and compliance with security constraints have been shown. This study lays the way for the development of more effective, sustainable, and resilient energy infrastructure by contributing to the advancement of the area of microgrid optimization.

Putting proteins in context

Proteins exhibit cell-type-specific functions and interactions, yet most ways of representing proteins lack any biological or environmental context. To address this gap, recent work by Li etal.1 introduces PINNACLE, ageometric deep learning approach that generates contextualized representations of proteins by combined analysis of protein interactions and multiorgan single-cell transcriptomics.

Advancing Iot Security: A Novel Intrusion Detection System for Evolving Threats In Industry 4.0 Using Optimized Convolutional Sparse Ficks Law Graph Point Trans-Net

With the rapid advancement of Industry 4.0, the integration of Internet of Things (IoT) strategies in industrial environments has increased exponentially. While this integration enhances productivity and efficiency, it also introduces significant security vulnerabilities. Previous research has employed several deep learning approaches for intrusion detection; however, these methods often suffer from insufficient accuracy, increased computational time, complexity, and higher error rates. To address these issues, this work proposes an innovative solution: "Advancing IoT Security: A Novel Intrusion Detection System (IDS) for Evolving Threats in Industry 4.0 using optimized Convolutional Sparse Fick's Law Graph Pointtrans-Net (CSFLGPtrans-Net)." The proposed system utilizes a comprehensive intrusion detection dataset composed of four different datasets: ToN-IoT, NSL-KDD, CSE‑CIC‑IDS2018, and IoT_bot. Initially, the input data undergoes a pre-processing stage that includes cleaning columns and rows, encoding features, and normalizing data. Following this, a hybrid optimization method, combining the Fire Hawk Optimizer with the Spider Wasp Optimizer, is applied for feature selection. This step is crucial for identifying the most significant features to enhance classification accuracy. The refined data is then classified using the CSFLGPtrans-Net model. To ensure secure data transfer, Fuzzy-based Elliptic Curve Cryptography (FECC) is employed. Experimental simulations conducted on the Python platform demonstrate that the proposed method outperforms existing approaches across various performance metrics, achieving a higher accuracy of 98% and a recall of 0.993. These results highlight the method's superior efficiency and potential for further advancement in securing Industry 4.0 environments.

EMNet: An ensemble deep learning approach for geological condition detection in tunnel excavation

This paper established a novel ensemble deep learning method for the identification of the geological condition encountered during tunnel boring machine (TBM) operation. By integrating multiple MobileNet base models with information fusion through the Dempster-Shafer theory (DST), the proposed method is able to provide reliable identification of the geological condition based on the images of the excavated mucks on TBM’s conveyor belt. Shapley additive explanations (SHAP) analysis is further presented for model interpretation and understanding of the key features in the images. A case study using data collected from Singapore’s Circle Line 6 tunnel construction projects is conducted to verify the effectiveness of the proposed method. The results indicate that (1) The proposed method detects the encountered geological condition with high accuracy; (2) EMNet outperforms all the base MobileNet models; (3) SHAP analysis demonstrates that surface texture could be the key feature that assists the model for the classification. The developed method contributes to the state of knowledge in proposing a novel ensemble method for reliable image identification and the state of practice in proposing a useful tool to identify the encountered geological condition automatically.

Automated Identification of Clinically Relevant Regions in Glaucoma OCT Reports Using Expert Eye Tracking Data and Deep Learning.

To propose a deep learning-based approach for predicting the most-fixated regions on optical coherence tomography (OCT) reports using eye tracking data of ophthalmologists, assisting them in finding medically salient image regions. We collected eye tracking data of ophthalmology residents, fellows, and faculty as they viewed OCT reports to detect glaucoma. We used a U-Net model as the deep learning backbone and quantized eye tracking coordinates by dividing the input report into an 11 × 11 grid. The model was trained to predict the grids on which fixations would land in unseen OCT reports. We investigated the contribution of different variables, including the viewer's level of expertise, model architecture, and number of eye gaze patterns included in training. Our approach predicted most-fixated regions in OCT reports with precision of 0.723, recall of 0.562, and f1-score of 0.609. We found that using a grid-based eye tracking structure enabled efficient training and using a U-Net backbone led to the best performance. Our approach has the potential to assist ophthalmologists in diagnosing glaucoma by predicting the most medically salient regions on OCT reports. Our study suggests the value of eye tracking in guiding deep learning algorithms toward informative regions when experts may not be accessible. By suggesting important OCT report regions for a glaucoma diagnosis, our model could aid in medical education and serve as a precursor for self-supervised deep learning approaches to expedite early detection of irreversible vision loss owing to glaucoma.

Enhancing the Recognition of Handwritten Arabic Characters through Hybrid Convolutional and Bidirectional Recurrent Neural Network Models

Handwritten Arabic character recognition remains a challenging task in pattern recognition due to the inherent complexities of the cursive script and visual similarities between characters. While deep learning techniques have demonstrated promising results in this domain, further enhancements to the model architecture can drive even greater performance improvements. This study introduces a hybrid deep learning approach that combines Convolutional Neural Networks (CNNs) with Bidirectional Recurrent Neural Networks, specifically Bidirectional Long Short-Term Memory (Bi-LSTM) and Bidirectional Gated Recurrent Units (Bi-GRU). By leveraging the strengths of both convolutional and recurrent neural network components, the proposed models are able to effectively capture spatial features as well as model the temporal dynamics and contextual relationships present in handwritten Arabic text. Experiments conducted on the AHCD and Hijjaa benchmark datasets show that the CNN-Bi-GRU framework achieved state-of-the-art accuracy rates of 97.05% and 91.78% respectively, outperforming previous deep learning-based methods. These results demonstrate the significant performance gains that can be achieved by integrating specialized temporal modeling and contextual representation capabilities into the handwriting recognition pipeline, without the need for explicit segmentation. The findings of this research represent a crucial advancement in the continued development of sophisticated and precise deep learning systems for Arabic handwriting recognition, with broad applications across domains that rely on efficient text extraction from handwritten documents.

Deep learning approaches for assessing pediatric sleep apnea severity through SpO2 signals

Pediatric Sleep Apnea–Hypopnea (SAH) presents a significant health challenge, particularly in diagnostic contexts, where conventional Polysomnography (PSG) testing, although effective, can be distressing for children. Addressing this, our research proposes a less invasive method to assess pediatric SAH severity by analyzing blood oxygen saturation (SpO2) signals. We adopted two advanced deep learning architectures, namely ResNet-based and attention-augmented hybrid CNN-BiGRU models, to process SpO2 signals in a one-dimensional (1D) format for Apnea–Hypopnea Index (AHI) estimation in pediatric subjects. Employing the CHAT dataset, which includes 844 SpO2 signals, the data was partitioned into training (60%), testing (30%), and validation (10%) sets. A predefined validation subset was randomly selected to ensure the models' robustness via a threefold cross-validation approach. Comparative analysis revealed that while the ResNet model attained an average accuracy of 72.9% across four SAH severity categories with a kappa score of 0.57, the CNN-BiGRU-Attention model demonstrated superior performance, achieving an average accuracy of 75.95% and a kappa score of 0.63. This distinction underscores our method's efficacy in both estimating AHI and categorizing SAH severity levels with notable precision. Further, to evaluate diagnostic capabilities, the models were benchmarked against common AHI thresholds (1, 5, and 10 events/hour) in each test fold, affirming their effectiveness in identifying pediatric SAH. This study marks a significant advance in the field, offering a non-invasive, child-friendly alternative for pediatric SAH diagnosis. Although challenges persist in accurately estimating AHI, particularly in severe cases, our findings represent a critical stride towards improving diagnostic processes in pediatric SAH.

Advanced drone-based weed detection using feature-enriched deep learning approach

This research addresses the pressing challenge of weed identification in agriculture, crucial for ensuring food security in anticipation of a global population exceeding 9.7 billion by 2050. Utilizing drone imagery, we collected a dataset and proposed a customized model to achieve optimal performance. Our proposed model uses strategically modified backbone, neck, and head components, leveraging elements such as Ghost Convolution, BottleNeckCSP, and ECA (Efficient Channel Attention) layers. These modifications enhance the model’s capability to discern intricate patterns in drone imagery, ultimately leading to improved precision in weed detection. We introduce a purposefully crafted dataset to complement the model’s training, and our experiments demonstrate superior performance compared to the baseline models. Our model achieves a precision of 72.5%, recall of 68.0%, and mAP@0.5 of 73.9, showcasing the effectiveness of our approach in enhancing detection accuracy. Leveraging a unique blend of feature extraction mechanisms, our model achieves remarkable accuracy in real-time soybean detection, outperforming established models like RT-DETR (Real-Time DEtection TransfoRmer) and YOLOv10. A detailed ablation study and comparative analysis with different YOLO versions and the transformer-based RT-DETR showcase the effectiveness of the proposed enhancements. Our work signifies a significant step toward advancing the field of precision agriculture, offering a model that is not only adaptive but also robust in identifying and localizing weeds in soybean fields.

A Novel Deep Learning Approach for Analyzing Glomerular Basement Membrane Lesions in a Mouse Model of X-Linked Alport Syndrome

Alport syndrome is a rare kidney disease typically more severe in males due to its X-linked inheritance. However, female patients with heterozygous X-linked Alport syndrome (XLAS) can develop renal failure over time, necessitating accurate pathological assessment for effective therapy. A key pathological finding in female XLAS patients is the mosaic pattern of partial loss of α5 chains of type IV collagen. This study, using a mouse model of XLAS with a nonsense mutation (R471*) in the Col4a5 gene, analogous to human XLAS, aimed to examine the consistency of this pattern with the glomerular basement membrane (GBM) structure. A modified periodic acid-methenamine silver (PAMS) staining method was developed for clearer GBM visualization. The integrated images from COL4α5-stained fluorescence, PAMS, and low-vacuum scanning electron microscopy (LVSEM) into a single-slide section and applied supervised deep learning to predict GBM lesions. Results showed significant individual variability in urinary protein levels and histological lesions. Pathological parameters, including crescent formation, focal segmental glomerulosclerosis, and the COL4α5/α2 ratio, correlated with clinical parameters like urinary protein and plasma creatinine levels. Integrated LVSEM analysis revealed dense GBM regions corresponded to areas where COL4α5 was preserved, while coarse GBM (basket-weave lesions) occurred in COL4α5-deficient regions. These advanced techniques can enhance biopsy-based diagnosis of Alport syndrome and aid in developing AI diagnostic tools for diseases involving basement membrane lesions.

Fully automated method for three-dimensional segmentation and fine classification of mixed dentition in cone-beam computed tomography using deep learning

ObjectiveTo establish a high-precision, automated model using deep learning for the fine classification and three-dimensional (3D) segmentation of mixed dentition in cone-beam computed tomography (CBCT) images. MethodsA high-precision, automated deep learning model was built based on modified nnU-Net and U-Net networks and was used to classify and segment mixed dentition. It was trained on a series of 336 CBCT scans and tested using 120 mixed dentition CBCT scans from three centers and 143 permanent dentition CBCT scans from a public dataset. The diagnostic performance of the model was assessed and compared with those of two observers with different seniority levels. ResultsThe model achieved accurate classification and segmentation of specific tooth positions in the internal and external mixed dentition datasets (Dice similarity coefficient: 0.964 vs. 0.951; Jaccard coefficient: 0.931 vs. 0.921; precision: 0.963 vs. 0.945; recall: 0.945 vs. 0.941; F-1 score: 0.954 vs. 0.943). These indices consistently exceeded 0.9 across multiple conditions, including fillings, malocclusion, and supernumerary tooth, with an average symmetric surface distance of 0.091 ± 0.029 mm. For permanent dentition, the Dice similarity and Jaccard coefficients exceeded 0.90, the average symmetric surface distance was 0.190 ± 0.092 mm, and precision and recall exceeded 0.94. With the aid of the model, the performance of junior dentists in mixed dentition classification and segmentation improved significantly; in contrast, there was no significant improvement in the performance of senior dentists. The speed of segmentation conducted by the dentists increased by 20.9–22.8 times. ConclusionThe artificial intelligence model has strong clinical applicability, robustness, and generalizability for mixed and permanent dentition. Clinical SignificanceThe precise classification and 3D segmentation of mixed dentition in dentofacial deformities, supernumerary teeth, and metal artifacts present challenges. This study developed a deep learning approach to analyze CBCT scans, enhancing diagnostic accuracy and efficacy. It facilitates detailed measurements of tooth morphology and movement as well as informed orthodontic planning and orthotic design. Additionally, this method supports dental education by assisting doctors in explaining CBCT images to the families of pediatric patients.

DeepCSNet: a deep learning method for predicting electron-impact doubly differential ionization cross sections

Abstract Electron-impact ionization cross sections of atoms and molecules are essential for plasma modelling. However, experimentally determining the absolute cross sections is not easy, and ab initio calculations become computationally prohibitive as molecular complexity increases. Existing AI-based prediction methods suffer from limited data availability and poor generalization. To address these issues, we propose DeepCSNet, a deep learning approach designed to predict electron-impact ionization cross sections using limited training data. We present two configurations of DeepCSNet: one tailored for specific molecules and another for various molecules. Both configurations can typically achieve a relative L2 error less than 5%. The present numerical results, focusing on electron-impact doubly differential ionization cross sections, demonstrate DeepCSNet's generalization ability, predicting cross sections across a wide range of energies and incident angles. Additionally, DeepCSNet shows promising results in predicting cross sections for molecules not included in the training set, even large molecules with more than 10 constituent atoms, highlighting its potential for practical applications.

Data-driven two-dimensional stationary quantum droplets and wave propagations in the amended Gross–Pitaevskii equation with two potentials via deep neural networks learning

In this paper, we develop a systematic deep learning approach to solve two-dimensional stationary quantum droplets (QDs) and investigate their wave propagation in the two-dimensional amended Gross–Pitaevskii equation with Lee–Huang–Yang correction and two kinds of potentials. Firstly, we use the initial-value iterative neural network algorithm for two-dimensional stationary QDs of stationary equations. Then, the learned stationary QDs are used as the initial value conditions for physics-informed neural networks to explore their evolutions in the space-time region. Especially, we consider two types of potentials, one is the two-dimensional quadruple-well Gaussian potential and the other is the P T -symmetric harmonic-oscillator (HO)-Gaussian potential, which lead to spontaneous symmetry breaking and the generation of multi-component QDs. The used deep learning method can also be applied to study wave propagations of other nonlinear physical models.

Influence of urban functional zone change on land surface temperature using multi-source geospatial data: A case study in Nanjing City, China

The urban heat island effect is a prevalent urban concern, and studies on its influencing factors are essential to mitigate its negative effects and promote sustainable urban development. Although many studies have revealed the impact of land cover change on land surface temperature (LST) during urbanization, most have focused only on built-up areas without considering the relationship between urban internal functional structure and LST. To clarify the influence of changes in urban functional zones (UFZs) on LST, focusing on the downtown area of Nanjing City, China, this study proposed an effective framework using multi-source geospatial data for qualitative and quantitative analysis. Specifically, very high spatial resolution images and point of interest data were jointly used to generate UFZ maps using an advanced deep learning approach, and LST distribution maps were retrieved from Landsat 8 images using the radiative transfer equation. The UFZ and LST maps were then overlaid and analyzed to reveal the impact of UFZ change on LST. The results showed that in 2013, 2018, and 2022, the residential region constituted the largest proportion of area, comprising 23.52%, 24.52%, and 27.28%, respectively. The areas of unused region and transport region experienced the most rapid decrease and increase of 6.08% and 4.05% from 2013 to 2022, respectively. Furthermore, the area of sub-high temperature zone decreased the most, by 3.47%, whereas that of the high-temperature zone increased the most, by 6.05%. In the study area, water and green space maintained relatively low LSTs, whereas residential region and unused region exhibited higher LSTs. From 2013 to 2022, UFZ changes contributed to a 0.71°C increase in LST. Transfers from green space, water, commercial region, transport region, and industrial region mostly resulted in a higher LST, which was associated with a decrease in fractional vegetation cover (FVC). Transfers from unused region, residential region, and public service region mostly contributed to reducing LST, which was caused by the increase in FVC. The findings and the proposed framework can provide reliable references for urban managers to optimize urban spatial layout and promote sustainable urban development.

Application of multimodal deep learning and multi-instance learning fusion techniques in predicting STN-DBS outcomes for Parkinson's disease patients

Parkinson's Disease (PD) is a progressive neurodegenerative disorder with substantial impact on patients' quality of life. Subthalamic nucleus deep brain stimulation (STN-DBS) is an effective treatment for advanced PD, but patient responses vary, necessitating predictive models for personalized care. Recent advancements in medical imaging and machine learning offer opportunities to enhance predictive accuracy, particularly through deep learning and multi-instance learning (MIL) techniques. This retrospective study included 127 PD patients undergoing STN-DBS. Medical records and imaging data were collected, and patients were categorized based on treatment outcomes. Advanced segmentation models were trained for automated region of interest (ROI) delineation. A novel 2.5D deep learning approach incorporating multi-slice representation was developed to extract detailed ROI features. Multi-instance learning fusion techniques integrated predictions across multiple slices, combining radiomics and deep learning features to enhance model performance. Various machine learning algorithms were evaluated, and model robustness was assessed using cross-validation and hyperparameter optimization. The MIL model achieved an area under the curve (AUC) of 0.846 for predicting STN-DBS outcomes, surpassing the radiomics model's AUC of 0.825. Integration of MIL and radiomics features in the DLRad model further improved discriminative ability to an AUC of 0.871. Calibration tests showed good model reliability, and decision curve analysis demonstrated clinical utility, affirming the model's predictive advantage. This study demonstrates the efficacy of integrating MIL, radiomics, and deep learning techniques to predict STN-DBS outcomes in PD patients. The multimodal fusion approach enhances predictive accuracy, supporting personalized treatment planning and advancing patient care.

Advances in vehicle re-identification techniques: A survey

The development of vehicle re-identification technology has significantly enhanced the operational efficiency of intelligent transport systems and smart cities, attributed to the advancement of artificial intelligence technologies such as deep learning and transformer models. By accurately tracking and identify the same vehicle under different cameras, the technology not only greatly enhances the ability of urban safety monitoring, traffic management and accident investigation, but also provides powerful technical support for the development of intelligent transportation. This paper explores the shift from traditional to deep learning approaches in vehicle re-identification, highlighting the rise of Transformer models. We assess both non-visual and vision-based re-identification technologies, with a special focus on the deep feature-based methods across supervised, unsupervised, and semi-supervised learning. And we summarize the performance of supervised and unsupervised methods on the VeRi-776 and VehicleID datasets. Finally, this paper outlines six directions for the future development of vehicle Re-ID technology, highlighting its potential applications in various areas such as smart city traffic management.

Effect of childhood atropine treatment on adult choroidal thickness using sequential deep learning-enabled segmentation

PurposeTo describe choroidal thickness measurements using a sequential deep learning segmentation in adults who received childhood atropine treatment for myopia control. DesignProspective, observational study. MethodsChoroidal thickness was measured by swept-source optical coherence tomography in adults who received childhood atropine, segmented using a sequential deep learning approach. ResultsOf 422 eyes, 94 (22.3 %) had no previous exposure to atropine treatment, while 328 (77.7 %) had received topical atropine during childhood. After adjusting for age, sex, and axial length, childhood atropine exposure was associated with a thicker choroid by 32.1 μm (95 % CI, 9.2–55.0; P = 0.006) in the inner inferior, 23.5 μm (95 % CI, 1.9–45.1; P = 0.03) in the outer inferior, 21.8 μm (95 % CI, 0.76–42.9; P = 0.04) in the inner nasal, and 21.8 μm (95 % CI, 2.6–41.0; P = 0.03) in the outer nasal. Multivariable analysis, adjusted for age, sex, atropine use, and axial length, showed an independent association between central subfield choroidal thickness and the incidence of tessellated fundus (P < 0.001; OR, 0.97; 95 % CI, 0.96–0.98). ConclusionsThis study demonstrated that short-term (2–4 years) atropine treatment during childhood was associated with an increase in choroidal thickness of 20–40 μm in adulthood (10–20 years later), after adjusting for age, sex, and axial length. We also observed an independent association between eyes with thicker central choroidal measurements and reduced incidence of tessellated fundus. Our study suggests that childhood exposure to atropine treatment may affect choroidal thickness in adulthood.

An end-to-end bi-objective approach to deep graph partitioning

Graphs are ubiquitous in real-world applications, such as computation graphs and social networks. Partitioning large graphs into smaller, balanced partitions is often essential, with the bi-objective graph partitioning problem aiming to minimize both the “cut” across partitions and the imbalance in partition sizes. However, existing heuristic methods face scalability challenges or overlook partition balance, leading to suboptimal results. Recent deep learning approaches, while promising, typically focus only on node-level features and lack a truly end-to-end framework, resulting in limited performance. In this paper, we introduce a novel method based on graph neural networks (GNNs) that leverages multilevel graph features and addresses the problem end-to-end through a bi-objective formulation. Our approach explores node-, local-, and global-level features, and introduces a well-bounded bi-objective function that minimizes the cut while ensuring partition-wise balance across all partitions. Additionally, we propose a GNN-based deep model incorporating a Hardmax operator, allowing the model to optimize partitions in a fully end-to-end manner. Experimental results on 12 datasets across various applications and scales demonstrate that our method significantly improves both partitioning quality and scalability compared to existing bi-objective and deep graph partitioning baselines.

RUL forecasting for wind turbine predictive maintenance based on deep learning

Predictive maintenance (PdM) is increasingly pursued to reduce wind farm operation and maintenance costs by accurately predicting the remaining useful life (RUL) and strategically scheduling maintenance. However, the remoteness of wind farms often renders current methodologies ineffective, as they fail to provide a sufficiently reliable advance time window for maintenance planning, limiting PdM's practicality. This study introduces a novel deep learning (DL) methodology for future-RUL forecasting. By employing a multi-parametric attention-based DL approach that bypasses feature engineering, thereby minimizing the risk of human error, two models—ForeNet-2d and ForeNet-3d—are proposed. These models successfully forecast the RUL for seven multifaceted wind turbine (WT) failures with a 2-week forecast window. The most precise forecast deviated by only 10 minutes from the actual RUL, while the least accurate prediction deviated by 1.8 days, with most predictions being off by only a few hours. This methodology offers a substantial time frame to access remote WTs and perform necessary maintenance, thereby enabling the practical implementation of PdM.