Search Results (3,999)

Background: Artificial intelligence (AI) is increasingly applied in pulmonary endoscopy, including diagnostic bronchoscopy, interventional pulmonology and endobronchial imaging. Advances in computer vision, machine learning and robotic systems have expanded the potential for automated lesion detection, navigation to peripheral pulmonary lesions, and real-time procedural support. However, the current evidence base remains heterogeneous, and translational challenges persist. Methods: This review summarizes current applications and developments of AI across white-light bronchoscopy (WLB), image-enhanced bronchoscopy (e.g., narrow-band imaging and autofluorescence imaging), endobronchial ultrasound (EBUS), virtual and robotic bronchoscopies, and workflow optimization and training. The authors also examine the methodological limitations, regulatory considerations, and implementation barriers that affect translation into routine practice. Results: Reported developments include deep learning-based models for mucosal abnormality detection, lymph-node characterization during EBUS-guided transbronchial needle aspiration (EBUS-TBNA), improved lesion localization, and reduction in operator-dependent variability. Additionally, AI-assisted simulation platforms and decision-support tools are reshaping training paradigms. Nevertheless, most studies remain retrospective or single-center, with limited external validation, dataset heterogeneity, unclear model explainability, and incomplete integration into clinical workflows. Conclusions: AI has the potential to support lesion detection, navigation, and training in pulmonary endoscopy. However, robust prospective validation, standardized datasets, transparent model reporting, robust data governance, multidisciplinary collaboration, and careful integration into clinical practice are required before widespread adoption. Full article

Open-pit mining operations rely heavily on visual inspection to identify indicators of slope instability such as surface cracks. Early identification of these geotechnical hazards enables timely safety interventions to protect both workers and assets in the event of slope failures or landslides. While computer vision (CV) approaches offer a promising avenue for autonomous crack detection, their effectiveness remains constrained by the scarcity of labelled geotechnical datasets. Deep learning (DL)-based models, in particular, require large amounts of representative training data to generalize to unseen conditions; however, collecting such data from operational mine sites is limited by safety, cost, and data confidentiality constraints. To address this challenge, this study proposes a novel hybrid game engine–generative artificial intelligence (AI) framework for large-scale dataset generation without requiring real-world training data. Leveraging a parameterized virtual environment developed in Unreal Engine 5 (UE5), the framework generates realistic images of open-pit surface cracks and enhances their fidelity and diversity using StyleGAN2-ADA. The synthesized datasets were used to train the YOLOv11 real-time object detection model and evaluated on a held-out real-world dataset of open-pit slope imagery to assess the effectiveness of the proposed framework in improving model generalizability under extreme data scarcity. Experimental results demonstrated that models trained using the proposed framework consistently outperformed the UE5 baseline, with average precision (AP) at intersection over union (IoU) thresholds of 0.5 and [0.5:0.95] increasing from 0.792 to 0.922 (+16.4%) and 0.536 to 0.722 (+34.7%), respectively, across the best-performing configurations. These findings demonstrate the effectiveness of hybrid generative AI frameworks in mitigating data scarcity in CV applications and supporting the development of scalable automated slope monitoring systems for improved worker safety and operational efficiency in open-pit mining. Full article

This systematic review critically examines the use of 2D and 3D digital imaging technologies of the face and teeth, with and without integration of artificial intelligence, for human identification in forensic and medicolegal contexts. Following PRISMA 2020 guidelines, Scopus, PubMed and Web of Science were systematically searched, identifying 26 studies published between 2011 and 2025 that met predefined eligibility criteria framed by a PECO-style question. Eighteen studies focused on facial imaging, six on dental imaging and two on integrated orofacial workflows, using digital photography, CCTV/video, 3D surface imaging, intraoral scanners, and three-dimensional superimposition methods, sometimes combined with classical algorithms and deep learning models. In controlled or semi-controlled settings, state-of-the-art facial algorithms often reported very high accuracy, with values up to 99.85%. By contrast, studies using real CCTV or other challenging forensic imagery showed more variable performance, with accuracies ranging from about 72.8% to 96.6%. Dental and orofacial studies reported 100% correct identifications for 3D superimposition of intraoral scans in small samples, and around 83% accuracy for automated AI-based dental identification. Crucially, fulfilling the promise of a true orofacial approach, this review proposes a structured research agenda focused on creating realistic multi-modal databases, standardizing protocols, and implementing probabilistic reporting (likelihood ratios) to guide future validation and legal admissibility. Full article

Background: It is significant to predict and fill metabolic reaction gaps (gap-fillings) for reconstructions of high-quality genome-scale metabolic models (GEMs). Currently, many existing optimization-based gap-filling methods have to rely on phenotypic data, while performances of topology-based approaches by deep learning algorithms need to be further improved. Methods: This paper proposes a novel topology-based approach (GHCN-SE) of predicting confidence scores of candidate reactions, which can be used for gap-fillings of GEMs. The topological features of GEMs are fully extracted by simultaneously using graph and hypergraph convolutional networks, such that both associations of metabolites in the same reaction and higher-order interactions of metabolites within reactions can be captured. After the feature fusion, we further employ the squeeze-and-excitation network to enhance features. Results: The reaction prediction and reaction recovery experiments through 5-fold cross validations on 108 high-quality BiGG GEMs show that the proposed GHCN-SE is superior to other related methods. The ablation study further demonstrates the contributions of the graph convolutional network, hypergraph convolutional network, and squeeze-and-excitation network in GHCN-SE. In addition, the visualization study interprets the effectiveness of GHCN-SE. Conclusions: For potential applications in metabolic engineering, biomedicine, etc., this proposed GHCN-SE can be used to further improve the phenotypic prediction accuracy of the draft GEM generated from automated reconstruction tools. Full article

Tumor budding (TB), defined as isolated single cells or small clusters of up to four tumor cells at the invasive front of colorectal carcinoma (CRC), is recognized as an important histopathologic marker associated with adverse tumor behavior. This review summarizes current knowledge on the morphologic assessment, biological significance, and clinical relevance of TB, emphasizing emerging artificial intelligence (AI) methods that aim to automate and standardize its quantification. Standardized reporting by the International Tumor Budding Consensus Conference (ITBCC) has improved reproducibility, while novel deep-learning algorithms demonstrate potential for objective and prognostically relevant TB assessment. Integration of AI-based TB evaluation with molecular and stromal biomarkers may refine patient stratification and facilitate personalized treatment strategies. Full article

The agricultural sector is undergoing a profound digital transformation driven by artificial intelligence, the Internet of Things, remote sensing, robotics, blockchain, and edge computing, which are being integrated into crop monitoring, irrigation management, disease detection, and supply chain transparency systems. This study employs systematic evidence mapping to characterize the applications of emerging digital technologies in sustainable agriculture; it delineates technological trajectories, areas of application, implementation gaps, and opportunities for improvement. Adhering to the PRISMA 2020 reporting protocol, 101 peer-reviewed articles indexed in Scopus and Web of Science (2020–2025) were identified, screened, and subjected to integrated thematic and bibliometric synthesis, using RStudio Version: 2026.01.1+403 and VOSviewer 1.6.20 for data mining on keywords and technological evolution patterns. Results show that deep learning and computer vision models achieved diagnostic accuracies of 90–99%, smart irrigation systems reduced water consumption by 10–30%, predictive yield models frequently reported R² values above 0.80, and greenhouse automation reduced energy consumption by approximately 20–30%. Blockchain-based architectures improved traceability and secure data transmission by 15–20%, while remote sensing integration enhanced spatial estimation accuracy up to R² = 0.92. The findings demonstrate a measurable transition toward data-driven, resource-efficient agricultural ecosystems supported by validated digital architectures. However, interoperability limitations, lack of standardized performance metrics, scalability challenges, and uneven geographical implementation—identified in nearly 40% of studies—highlight the need for harmonized evaluation frameworks, cross-platform integration standards, and long-term field validation to ensure sustainable and scalable digital transformation. Full article

Labor shortages and the need for increased productivity have accelerated the development of robotic harvesting systems for greenhouse crops; however, reliable operation under fruit occlusion and clustered arrangements remains a major challenge, particularly due to the limited integration between perception and motion planning modules. The paper presents the design and experimental validation of an autonomous robotic system for greenhouse tomato harvesting. The proposed platform integrates a rail-guided mobile base, a six-degrees-of-freedom robotic manipulator, and an adaptive end effector with a hybrid vision framework that combines convolutional neural networks and watershed-based segmentation to enable robust fruit detection and localization under occluded conditions. The proposed approach enables improved separation of overlapping fruits and provides accurate spatial localization through stereo vision combined with IMU-assisted camera-to-robot coordinate transformation. An occlusion-aware trajectory planning strategy was developed to generate collision-free manipulation paths in the presence of leaves and stems, enhancing harvesting safety and reliability. The system was trained and evaluated using a dataset of real greenhouse images supplemented with synthetic data augmentation. Experimental trials conducted under practical greenhouse conditions demonstrated a fruit detection precision of 96.9%, recall of 93.5%, and mean Intersection-over-Union of 79.2%. The robotic platform achieved an overall harvesting success rate of 78.5%, reaching 85% for unobstructed fruits, with an average cycle time of 15 s per fruit in direct harvesting scenarios. The rail-guided mobility significantly improved positioning stability and repeatability during manipulation compared with fully mobile platforms. The results confirm that integrating hybrid perception with occlusion-aware motion planning can substantially improve the functionality of robotic harvesting systems in protected cultivation environments. The proposed solution contributes to the advancement of automation technologies for greenhouse vegetable production and supports the transition toward more sustainable and labor-efficient agricultural practices. Full article

Deep learning is increasingly integrated into oral rehabilitation workflows, particularly in implant planning, prosthodontic design automation, and peri-implant diagnosis. However, reported performance is heterogeneous and difficult to compare across tasks, modalities, and validation designs. The goal of this study was to critically analyze deep learning architecture families applied to oral rehabilitation and to provide task-driven selection guidance supported by an evidence table reporting dataset characteristics, validation strategy, and performance metrics. A focused narrative review was conducted using transparent, database-specific search criteria (final n = 10 included studies), emphasizing implant planning (cone–beam computed tomography [CBCT]-based segmentation), prosthodontic design (intraoral scan [IOS]/mesh inputs), and peri-implant diagnosis (periapical/panoramic radiographs). Evidence certainty for each clinical task was assessed using GRADE-informed ratings (High/Moderate/Low/Very Low). Extracted variables included clinical task, imaging modality, dataset size, architecture, validation strategy (internal vs. internal + external), split level, ground truth protocol, and performance metrics. A structured computational and hardware feasibility analysis was conducted for each architecture family to support real-world deployment planning. Encoder–decoder networks (U-Net/nnU-Net) dominate CBCT segmentation for implant planning, while detection architectures (Faster R-CNN, YOLO) support implant localization and peri-implant assessment on radiographs. Generative models (3D GANs, transformer-based point-to-mesh networks) enable crown design from three-dimensional scans. Hybrid CNN–Transformer architectures show promise for multimodal CBCT–IOS fusion, though direct evidence from the included studies remains limited to a single study. External validation remains uncommon yet essential given the risk of domain shift. In conclusion, architecture selection should be anchored to task geometry (2D vs. 3D), artifact burden, and required clinical output type. Reporting standards should prioritize dataset transparency, validation rigor, multi-center external testing, and uncertainty-aware outputs. Full article

Acute Lymphoblastic Leukemia (ALL) is a highly prevalent hematological malignancy, especially in children, for whom precise and prompt subtype identification is essential to establish suitable treatment protocols. Current deep learning-based computer-aided diagnosis (CAD) methods for identifying ALL are hindered by numerous drawbacks, such as a dependence on solely spatial feature depictions, elevated feature dimensions, computationally extensive deep learning architectures, inadequate multi-layer feature utilization, and poor interpretability. This paper introduces LUMINA-Net, a custom, lightweight, and interpretable deep learning CAD for the automated identification and subtype diagnosis of ALL using microscopic blood smear pictures. LUMINA-Net makes four principal contributions: first, it integrates a self-attention module within a lightweight custom Convolution Neural Network (CNN) to effectively capture long-range spatial relationships across clinically pertinent cytological patterns while preserving a compact design. Second, it employs a Discrete Wavelet Transform (DWT)-based wavelet pooling layer that decreases feature dimensions by up to 96.875% while enhancing the obtained depictions with spatial-spectral information. Third, it utilizes a multi-layer feature fusion strategy that combines wavelet-pooled features from two deep layers with a third fully connected layer to create a discriminating multi-scale feature vector. Fourth, it incorporates Gradient-weighted Class Activation Mapping as a dedicated explainability process to furnish clinicians with apparent visual explanations for each classification decision. Withoit the need for image enhancement or segmentation preprocessing, LUMINA-Net outperforms the competing state-of-the-art methods on the same dataset, achieving a peak accuracy of 99.51%, specificity of 99.84%, and sensitivity of 99.51% on the publicly available Kaggle ALL dataset. This demonstrates that LUMINA-Net has the potential to be a dependable, effective, and clinically interpretable CAD tool for ALL diagnosis. Full article

Background: Pneumonia of unknown etiology (PUE), defined as pneumonia cases without an identified pathogen at the time of clinical presentation, represents a critical clinical warning signal for emerging infectious disease (EID) outbreaks with pandemic potential. Yet, conventional pathogen-centric surveillance systems suffer from an inherent blind spot: they cannot detect early clustering signals before the causative agent is identified, creating a window of vulnerability during novel pathogen emergence. To address this gap, this study aims to develop a deep learning model that leverages unstructured chest imaging text—a routinely available clinical data stream—to enable real-time, automated screening of PUE cases and early warning of EID clusters, independent of prior pathogen knowledge, within an integrated multi-pathogen surveillance framework. Methods: We retrospectively collected data from 8860 patients with respiratory illnesses at a tertiary hospital in Beijing, China, including 980 PUE cases (11.1%) and 7880 known-etiology pneumonia cases. A deep learning model (RoBERTa with attention enhancement) was developed using unstructured chest imaging reports. The Matthews correlation coefficient (MCC) curve was employed to determine the optimal decision threshold. Model performance was assessed for PUE case identification and clustering signal detection on a test set. Results: The model achieved an area under the receiver operating characteristic curve of 0.986 (95% CI: 0.981–0.991). At the optimal threshold of 0.08, selected by maximizing the Matthews correlation coefficient (MCC)—a balanced metric that accounts for all four confusion matrix outcomes—sensitivity was 89.8%, and specificity was 97.0% for identifying PUE cases. In a simulated surveillance exercise, the model showed a high correlation between the predicted and actual case counts (Pearson’s r = 0.901), suggesting its potential to detect abnormal clustering signals prior to pathogen identification. Conclusions: The developed model demonstrates potential to detect clustering signals of PUE caused by unknown pathogens and can be integrated with hospital information systems, providing a feasible, low-cost tool for integrated surveillance of pathogens with pandemic potential. This approach enables earlier outbreak detection and supports public health decision-making during the critical window before pathogen identification. Full article

Varroa destructor (Bee mite) and Deformed Wing Virus are primary causes of honeybee colony collapse. This study developed an automated AI-based vision inspection system for detecting bee mites and deformed bees using the YOLO algorithm. The system integrates an RGB camera, a beecomb rotation motor, and an image transmission module to enable automated dual-sided image acquisition of the beecomb. The image characteristics of normal bees, bee mites, and deformed bees were analyzed, and YOLO-based object detection models were developed to classify them. Six YOLO models—based on YOLOv8 and YOLOv11 architectures across three model sizes (nano, small, and large)—were evaluated on 405 test images (6441 objects). The proposed system reduced the inspection time from 240 s required for manual method to 20 s per beecomb, achieving 12-fold efficiency improvement. Comparative analysis showed model-task specialization: YOLOv8l excelled in detecting small bee mites (F1: 92.5%, mAP[0.5]: 92.1%), while YOLOv11s achieved the highest performance for morphologically diverse deformed bees (F1: 95.1%). Error analysis indicated that detection performance was influenced by morphological characteristics. Deformed bee detection errors correlated with overlap in wing-to-body ratio: DB Type II exhibited 18.6% miss rate, while DB Type III achieved perfect detection. In bee mite detection, a sensitivity–specificity trade-off was observed: YOLOv11l had the lowest false negatives (2.5%) but highest false positives, while YOLOv8l demonstrated superior discrimination. These results demonstrate the practical potential of the proposed system for field deployment in apiaries, supporting early pest diagnosis and improved colony health management. The model-task specialization framework provides guidance for architecture selection based on object characteristics. Future work will focus on multi-location validation and real-time monitoring integration. Full article

The industrial metaverse is a digital twin space that integrates the real world with virtual environments through bidirectional synchronization. It supports critical services, such as time-sensitive machine control and large-scale collaboration, which require Time-Sensitive Networking and scalable Data Distribution Services. DDS, developed by the Object Management Group, provides excellent scalability and diverse QoS policies but struggles to guarantee transmission delay and jitter for time-critical applications. TSN, based on IEEE 802.1 standards, addresses these challenges by ensuring time-criticality. However, current research lacks comprehensive integration mechanisms for DDS and TSN, particularly from the viewpoints of semantics and system framework. Additionally, there is no adaptive QoS mapping converting the abstract DDS QoS policies to the sophisticated TSN QoS parameters. This paper presents a novel DDS-over-TSN framework that incorporates three key functions to address these challenges. First, Cross-layer QoS Mapping automates correspondences between DDS and TSN parameters, deriving technical constraints from standard documentation through retrieval-augmented generation. Second, Semantic Priority Estimation extracts substantial priority levels by utilizing language model embedding vectors as high-dimensional feature extractors. Third, Adaptive Resource Allocation performs dynamic bandwidth distribution for each priority level through reinforcement learning. Simulation results reveal over 99% mapping accuracy and 97% consistency in priority extraction. The applied Deep Reinforcement Learning paradigm allocated 99% of required resources to high-priority classes and reduced resource wastage by 15% compared to conventional methods. This methodology meets industrial requirements by ensuring both deterministic real-time performance and efficient resource isolation. Full article

With the rapid development of e-commerce and smart manufacturing, automated warehouse systems have become critical infrastructure for modern logistics. In China’s vast market, the dynamic scheduling of Rail-Guided Vehicles (RGVs) faces significant challenges due to complex task uncertainties, hierarchical supply chain structures, and real-time collision avoidance requirements. Traditional rule-based methods and static optimization models often fail to adapt to such dynamic environments. To address these issues, this paper proposes a novel hybrid deep reinforcement learning framework integrating a Dynamic Graph Neural Network (DGNN) and a Transformer model. The DGNN captures the spatiotemporal dependencies of the warehouse network topology, while the Transformer mechanism enhances long-range feature extraction for task prioritization. Furthermore, we design a centralized Deep Q-network (DQN) framework with parameterized action spaces to coordinate multiple RGVs collaboratively. While the system manages multiple physical vehicles, the learning architecture employs a single-agent global scheduler to avoid the non-stationarity issues inherent in multi-agent reinforcement learning. Experimental results based on real-world data from a large-scale electronics manufacturing warehouse demonstrate that our method reduces average task completion time by 18.5% and improves system throughput by 22.3% compared to state-of-the-art baselines. The proposed approach demonstrates potential for intelligent warehouse management in dynamic industrial scenarios. Full article

Coral reef ecosystems are declining rapidly due to climate change, disease, and anthropogenic stressors, driving the expansion of land-based coral propagation for reef restoration. A major bottleneck in these efforts is the manual measurement of coral recruit tissue area from microscopy images, which requires 2–7 min per image and limits scalability. We present a hierarchical deep learning pipeline that automates this measurement by integrating YOLO-based detection with Segment Anything Model (SAM) segmentation. YOLO localizes recruits and classifies them by developmental stage; stage-specific fine-tuned SAM models then segment live tissue using bounding box and background point prompts to suppress segmentation leakage and improve boundary precision. Surface area is computed directly from the segmented masks using pixel size extracted from image metadata. The pipeline reduces processing time to approximately 3–5 s per image—a 24–140× speedup over manual tracing. Evaluated on 3668 microscopy images from two national coral research facilities, the system achieves a mean IoU exceeding 95% and an auto-acceptance rate (AAR) of 71.51%, where predicted-to-ground-truth area ratios fall within a ±5% tolerance of expert annotation, substantially reducing manual workload while maintaining measurement reliability across species, developmental stages, and imaging conditions. This workflow addresses a critical bottleneck in restoration research and demonstrates the broader applicability of AI-based image analysis in marine ecology. Full article

Deep learning has emerged as an effective tool for automatic waste classification, supporting cleaner cities and more sustainable recycling systems. Because environmental protection is central to the United Nations Sustainable Development Goals (SDGs), improving the sorting and processing of everyday waste is a practical step toward this broader objective. In many real-world settings, however, waste is still sorted manually, which is slow, labor-intensive, and prone to human error. Although convolutional neural networks (CNNs) can automate this task with high accuracy, many state-of-the-art models remain too large and computationally demanding for low-cost edge devices intended for deployment in homes, schools, and small recycling facilities. In this work, we investigate lightweight waste-classification models suitable for TinyML deployment while preserving competitive accuracy. We first benchmark multiple CNN architectures to establish a strong baseline, then apply complementary compression strategies including quantization, pruning, singular value decomposition (SVD) low-rank approximation, and knowledge distillation. In addition, we evaluate an RL-guided multi-teacher selection benchmark that adaptively chooses one teacher per minibatch during distillation to improve student training stability, achieving up to 85% accuracy with only 0.496 M parameters (FP32 ≈ 1.89 MB; INT8 ≈ 0.47 MB). Across all experiments, the best accuracy–size trade-off is obtained by combining knowledge distillation with post-training quantization, reducing the model footprint from approximately 16 MB to 281 KB while maintaining 82% accuracy. The resulting model is feasible for deployment on mobile applications and resource-constrained embedded devices based on model size and TensorFlow Lite Micro compatibility. Full article