Search Results (911)

Few-shot anomaly detectors for industrial inspection typically use between five and 20 reference images to learn what a normal part looks like. These images are assumed to be defect-free, which is not 100% guaranteed on an actual production line. This makes the reference set a potential point of failure: if an image of a defective part ends up among the references, the detector will silently accept similar defects. First, we quantify this degradation using 35 categories from the public datasets MVTec AD, VisA, BTAD, and MVTec LOCO. We show that replacing just three of 10 reference images with defective parts results in a loss of 1.8 percentage points of AUROC on average, and up to 3.0 depending on the dataset. This represents a substantial loss for an industrial anomaly detection system. Second, we present an unsupervised audit algorithm that identifies which reference images are likely defective, and requires no training. The audit is a one-time preprocessing step that evaluates the internal consistency of the reference set before it is used for inspection. Each image is compared to the rest of the references, scored based on its most atypical region, and flagged when that score stands out. A technician can then remove, review, or recapture the flagged images before the system is deployed on the production line. We show that recapture recovers more than half of the accuracy lost due to contamination, and if there are no truly defective images, it does not degrade the system’s performance. Our method is intended for memory-bank detectors built from a handful of references, where an isolated part with a structural or textural defect can slip into the set. It runs in seconds on CPU, makes its decisions at the image level independently of the anomaly detection algorithm used, and can be easily incorporated into existing systems. The code and results are publicly available. Full article

Automated inspection of small metallic components is increasingly required in industrial manufacturing, particularly when dimensional conformity and geometric characteristics affect assembly quality and functional performance. This paper presents a preliminary proof-of-concept study for a collaborative SCARA-based platform intended for automated dimensional and geometric inspection of metallic plates. The study does not report a complete metrologically validated inspection cell; instead, it defines the industrial inspection problem, proposes a modular robotic architecture, and documents the preliminary implementation activities already completed. The industrial motivation is associated with metallic plates used in automotive thermal-management components, while a simplified rectangular plate is adopted for controlled development of handling, positioning, image acquisition, calibration, and initial dimensional-feature extraction. The proposed architecture separates robotic manipulation from the measurement subsystem: the SCARA robot provides repeatable handling and positioning, whereas the vision system supports preliminary calibrated in-plane measurement. The robotic platform is based on the PSR-20 collaborative SCARA robot (SmileTech, Portugal), selected due to its compact footprint, repeatable planar motion, affordable cost and suitability for automated pick-and-place inspection tasks. The implemented activities include SCARA workspace simulation, end-effector design, camera integration, intrinsic camera calibration, measurement-plane calibration, and preliminary calibrated 2D dimensional measurements. The experimental validation reported in this study is currently limited to in-plane measurements of length and width. Height, thickness, flatness, and surface-form inspection are identified as future extensions and have not yet been experimentally validated. The results show repeatable measurements under the tested conditions for a high-contrast reference part, while the representative metallic part remains limited by segmentation robustness and contour-definition uncertainty. Remaining work includes a formal uncertainty analysis, broader validation with calibrated reference artefacts, robust segmentation, and out-of-plane sensing. Full article

Power transformers are critical, capital-intensive assets in electricity distribution networks, and their failure disrupts supply while incurring substantial repair costs. Building on an earlier probabilistic analysis of power transformers with six failure modes and inspection, this paper proposes a twelve-state semi-Markov extension that additionally captures three operational realities not previously modeled: condition-based degradation warnings from dissolved gas or partial discharge monitoring, a regime-switching failure rate driven by seasonal peak load conditions, and a repair structure that separates an immediate in-stock repair from a spare-part-wait state and a standby-covered repair state. Reliability indices—mean time between failures, steady-state availability, and the expected busy period of the repair crew—are derived under the special case of exponential resolution times: MTBF is obtained in closed form via a first-passage-time argument on the up-state cycle, while availability and busy period are obtained from the steady-state solution of the twelve-state generator matrix and evaluated numerically, together with a full sensitivity and relative sensitivity analysis of all thirteen model parameters and, building on that analysis, approximate confidence intervals for MTBF, availability, and busy period obtained by delta method propagation of parameter uncertainty. Because condition monitoring, seasonal exposure, and standby utilization records were not part of the original data collection, illustrative parameter values grounded in transformer loading standards and utility maintenance practice are used to demonstrate the model, and the sensitivity analysis quantifies which of these illustrative values would most change the results once replaced by field estimates. A dedicated ±50% what-if analysis confirms the headline conclusions are robust to substantial mis-specification of the two most sensitive parameters, and a distribution shape analysis further confirms, analytically, that the steady-state indices are invariant to the choice of sojourn time family (exponential, Weibull, or otherwise) once means are fixed, underscoring the genuinely semi-Markov (not merely CTMC) character of the model. The extension is positioned as a direct methodological successor to two prior studies of the same transformer fleet, and a data collection plan is given to convert the illustrative parameters into field-estimated ones. Full article

The impact of financial and accounting oversight depends in part on whether inspections deter corporate misdeeds. Using deterrence theory, we ask whether accounting inspections constrain corporate misconduct, which governance responses may explain the association, and under what conditions it is stronger. We examine the Ministry of Finance’s accounting information quality inspections using a difference-in-differences design with staggered inspection timing for Chinese A-share listed firms over 2007–2024. The baseline estimates show lower misconduct probability and frequency after inspection. Mechanism analyses examine internal control deficiencies, managerial myopia, and audit monitoring as possible explanations. The effects are more pronounced for less complex businesses, for regions with weaker legal environments, and for industries with more intense competition. The findings also differ depending on the type of misconduct, with more pronounced effects for serious misconduct and disclosure-related misconduct. These findings contribute to the literature on the relationship between accounting inspections and corporate misconduct. They also influence the inspection and remediation practices of fiscal authorities, the internal control improvements of firms and the risk assessments of auditors. Full article

Printed product surface defect inspection faces critical challenges in real industrial production, including insufficient detection accuracy, poor robustness, and scarcity of labeled defective samples. This paper proposes an improved Siamese network for high-resolution printed image defect detection to address the above limitations. Original 1300 × 460-pixel printed images are segmented into 64 × 64 patches to construct a training and testing dataset. We optimize the feature measurement strategy of contrastive loss and integrate a position matching module and multi-threshold evaluation strategy to balance detection precision and stability. Quantitative experiments demonstrate that the proposed method achieves an overall accuracy of 94.35%, a precision of 91.99%, a recall of 95.17%, and an F1-score of 93.55% at the fixed threshold of 0.45. Compared with mainstream lightweight detection models including YOLOv8Lite (You Only Look Once version 8 Lite) and MobileViTv2 (Mobile Vision Transformer version 2), our method maintains a stable F1-score advantage of 10.84–24.36% under optimal threshold settings. The proposed approach achieves outstanding performance and robustness in few-shot scenarios and can be deployed for automatic quality inspection of printed parts used in power communication equipment. Full article

A flexible part has no unique shape until it is constrained, which makes dimensional inspection difficult. Standard practice clamps it in a dedicated jig and probes it with a coordinate measuring machine or a range sensor. We replace the jig with virtual fixturing. From partial range views of the unfixtured part, the pipeline recovers a coarse pose between the scan and the CAD model using a robust geodesic bilateral curvature algorithm, deforms the model towards the scan by non-rigid registration, and computes deviations along the model surface normal to decide whether the part is in tolerance. On a prismatic part and a game-controller housing, the method is more accurate than optimal-step non-rigid ICP, radial basis function FEM, and coherent point drift, because the generated deformations come from an operator closely related to the proposed method’s own regularizer, those margins favor it by construction and are reported with that caveat. On a generated thin-shell test part the measurement uncertainty of the implementation is measured at about 0.039 mm, dominated by the registration rather than by the sensor. The pipeline is then exercised on a physically scanned injection-molded engine cover, a 612 mm part whose free-state residual against its nominal model is 4.62 mm at the verified global optimum. No independent coordinate-measuring-machine reference was available for that part, so this experiment is reported as a free-state residual and a controlled comparison with and without the feature set, not as a statement of absolute measurement accuracy. Full article

Software reliability is an important part of reliability assurance for complex engineering systems, and timely fault diagnosis supports safe and continuous operation. After a test failure, statement-level fault localization ranks source-code lines for early inspection. Representation-based approaches can operate without a coverage matrix by extracting line-level hidden states from a frozen large language model for code (code LLM), but their downstream readout and training objective are not necessarily aligned with source-window ranking and buggy-version-level evaluation. We propose SeqRankFL, a sequence-aware ranker for LLM-based code representations. It combines a bidirectional long short-term memory network (BiLSTM), which follows code order within a source window, with a Hybrid objective that joins binary cross-entropy (BCE) and the listwise learning-to-rank loss ListNet. On the source-window ranking task using windows of at most 128 physical lines within known buggy files from BugsInPy and Defects4J, relative to a matched LLMAO-style Transformer+BCE reference that shares the same representations, splits, and evaluation protocol, SeqRankFL raises the equally weighted Top-1 from 51.31% to 58.58%, an absolute gain of 7.27 percentage points, with consistent improvements on both datasets. Further controlled analyses show that multi-depth layer mixing and syntax scope mainly improve average ranks, whereas control-flow/data-flow graphs and failure behavior have condition-dependent effects across models, languages, and project partitions. The controlled results identify the listwise objective as the primary driver of the Top-1 improvement, with the BiLSTM readout providing an additional architecture-dependent gain under the ranking-aware objective. Full article

This paper presents the design and implementation of an automated quality inspection system based on rotary indexing and inline dimensional measurement. The proposed system integrates a rotary table for precise part positioning with a contact displacement sensor for high-resolution measurement, enabling real-time evaluation of critical geometric parameters during the production cycle. A PLC-based control architecture ensures synchronized operation between mechanical motion and measurement acquisition, utilizing industrial communication over PROFINET for reliable data exchange. A dedicated decision algorithm is implemented to classify parts as acceptable or defective based on predefined tolerance thresholds, allowing immediate feedback and process control. Experimental validation demonstrates high repeatability and measurement stability under continuous operation, confirming the suitability of the system for industrial applications requiring fast and reliable quality control. The results highlight the effectiveness of combining rotary indexing mechanisms with inline measurement techniques for improving production efficiency and reducing inspection time. Full article

Metal additive manufacturing (AM) enables compact hydraulic manifolds with curved internal channels, reduced part count and integrated functionality. However, these benefits also create major inspection challenges, especially in thick metallic sections containing closely spaced and intersecting passages. This study examines the practical use of X-ray computed tomography (CT) for an industrial hydraulic manifold manufactured from maraging steel MS1 by laser powder bed fusion (LPBF). Selected cross-sections from the reconstructed CT volume were compared with the nominal computer-aided design (CAD) geometry and evaluated using a qualitative interpretability classification supported by comparative image contrast-to-noise ratio (CNR) analysis and approximate projected steel thicknesses. Clearly interpretable regions showed higher and more consistent CNR, whereas geometrically congested regions generally exhibited lower and more variable local contrast. However, projected material thickness alone did not determine interpretability, indicating an additional influence of geometric overlap, orientation and reconstruction artefacts. Particular attention was given to a channel wall adjacent to a locally collapsed external support structure. No spatially persistent through-wall discontinuity was identified, although smaller defects, local wall thinning and metallurgical changes could not be excluded. A pneumatic immersion test at 0.8 MPa showed no visible bubble formation or observable pressure decrease. This pressure exceeded the expected operating pressure of the affected relief or tank channel but was substantially below the 35 MPa maximum intended pressure of the pressure-side circuits and therefore did not constitute structural qualification. The study demonstrates that whole-component CT can provide useful local inspection information for complex LPBF manifolds, but its reliability depends strongly on local geometry and acquisition conditions. Quantitative image assessment and complementary functional testing may therefore be required when CT results are insufficient for complete qualification. Full article

Bovine bacterial mastitis remains a major challenge for the dairy industry, and timely identification of causative bacteria is essential for effective disease management. This study compared six regions of the 16S rRNA gene for PCR high-resolution melt (HRM) curve analysis to identify the most suitable target for species-level differentiation of mastitis-associated bacteria. A total of 547 clinical bovine milk samples from cows with clinical mastitis were cultured, and representative isolates of 18 bacterial species were selected for comparative analysis. Species identity was confirmed using matrix-assisted laser desorption ionisation time-of-flight mass spectrometry (MALDI-TOF MS) prior to molecular analysis. All six primer pair combinations were tested across the 18 species panel, and discriminatory performance was assessed using a genotype confidence percentage (GCP)-based classification model that allowed for objective interpretation without visual inspection. Among the six target regions, the 203 bp amplicon generated by primer pair F1R1, spanning the V1-to-V2 region and part of V3, showed the strongest overall discriminatory performance with no cross-classification observed. Sequencing of F1R1 amplicons confirmed species identity and showed concordance with HRM-based classification. Phylogenetic analysis of sequences from this region showed clustering broadly consistent with established taxonomic relationships. Proof-of-concept testing showed that the workflow could also be applied to DNA extracted directly from selected milk samples without prior bacterial isolation. Overall, this approach provides a rapid, cost-effective, and high-throughput method for preliminary discrimination of mastitis-associated bacteria. Full article

Enterprise strategic decisions must reconcile conflicting stakeholder interests under time pressure, yet the consulting that traditionally supports them remains out of reach for most small and medium-sized enterprises. Large language models make parts of this work automatable, but a single model speaks with one voice, and its reasoning can be neither inspected nor contested. As frontier models continue to improve, whether structured multi-agent deliberation is worth its additional cost has therefore become an empirical question rather than a design assumption. This study designs, prototypes, and evaluates Servi.AI, a personality-aware multi-agent intelligent decision support system for enterprise strategy. The system grounds every recommendation in a traceable evidence chain retrieved over a knowledge graph. It stages a statement–discussion–consensus roundtable in which role-specialized agents argue from conflicting professional stances. It also simulates how synthetic stakeholders, calibrated against a public personality dataset of 874,434 respondents, will experience the candidate decision. The roundtable characterizes how a decision is argued, whereas the sandbox characterizes how it will be experienced. A questionnaire with 133 screened decision-makers confirms these requirement priorities. We evaluate the system across five experimental axes: 2800 controlled simulation runs and a twelve-case benchmark judged blind across three model families. A strong single model attains the highest holistic scores (8.22–8.56/10 across judges), while deliberation contributes auditable role-grounded reasoning, conflict surfacing, and an executable blueprint. Ablating retrieval loses all 71 valid pairwise comparisons, while a heterogeneous five-family agent pool significantly improves risk coverage (Cliff’s $\delta =+0.75$). Retrieval thus drives the evidence-side qualities, and the role structure drives the deliberation-side ones: deliberative value is decomposable along architectural components, a middle-range design proposition. These findings support selective rather than default deployment. The released benchmark, judging protocol, and raw results provide a reusable basis for deciding when multi-agent decision support is worth its cost. Full article

Spaceborne synthetic aperture radar (SAR) needs point-like reference targets for co-registration, calibration, and absolute geolocation, but real corner reflectors (CRs) exist only at a few dedicated sites. We present a detection and three-dimensional (3D) positioning pipeline for naturally occurring corner-reflector-like (pseudo-CR) scatterers in KOMPSAT-5 stacks, combining constant false alarm rate (CFAR) detection ranked by local contrast rather than absolute power, a multi-scale template bank, multi-look range-Doppler bundle triangulation, and a per-target conditioning diagnostic vector. At the nine-CR Mongolia calibration field, all nine are recovered, with a mean horizontal error of 1.37 m and a mean 3D error of 2.01 m. At a CR-absent suburban site in Suncheon, Republic of Korea, 18 scenes yield 39,822 clusters, 26,659 of which pass the conditioning gate. Over 16 response-enriched poles retained by author adjudication, the difference to the nearest conditioning-filtered cluster averages 3.19 m horizontally and 1.45 m vertically relative to the surveyed pole base. Because part of that set was selected after inspecting preliminary responses, these are conditional nearest-cluster agreements, not verified target accuracies, and no detection probability is claimed. In an exploratory comparison on the same set, contrast ranking retains more evaluation targets than power ranking at equal budgets, 16 against 9 of 16. Removing robust weighting and pruning degrades the mean Mongolia 3D error to 7.39 m; the DEM prior is non-critical, whereas the clustering radius is consequential in the dense pole field, with the count within 20 m falling from 16 to 9 as the radius grows from 15 m to 30 m. In a leave-scene-out check, 85.8% of 254 evaluable clusters persist in a scene excluded from their solutions, against 35.2% of 270 evaluable random positions: a complete-case enrichment of 78%, or 53–82% under cross-arm worst-case imputation of the unevaluable targets; neither estimate measures catalogue precision. Limited KOMPSAT-5 stacks can therefore yield few-metre nearest-cluster agreement at selected pole locations where stable multi-scene responses form. Full article

Deploying prohibited-item detectors on resource-constrained X-ray security inspection terminals is not equivalent to selecting the smallest available YOLO scale: excessive compression can reduce feature capacity, whereas medium-scale detectors may retain avoidable computational redundancy. This study therefore investigates a deployment-oriented operating point through coordinated backbone compression, feature compensation, and class-sensitive optimization. YOLO-SMV and YOLO-EMV are developed from YOLOv8m and YOLO11m by combining MobileNetV3-Small backbone reconstruction, SE/ECA-based channel recalibration, and a VF-BCE objective for difficult and underrepresented categories. A three-seed full-factorial study on YOLO11m shows that VF-BCE provides the largest individual accuracy gain and that ECA repeatedly recovers part of the performance lost in the compressed VF-BCE pathway. The accuracy-oriented ECA+VF-BCE configuration reaches $0.93842\pm 0.00720$ mAP50, whereas the deployment-oriented YOLO-EMV reaches $0.92744\pm 0.00768$; the latter trades $1.098\pm 0.061$ percentage points of mAP50 for a reduction from 20.03 M to 12.04 M parameters and from 67.9 G to 28.8 G FLOPs. Under the common seed-41 SIXray protocol, YOLO-EMV also achieves higher mAP50 than standard YOLO11n, YOLO11s, and YOLO11m, demonstrating that the selected operating point is not reproduced simply by choosing a smaller baseline. Published SIXray results are reported separately as protocol-aware literature context rather than as a cross-paper ranking. Additional OPIXray and PIDray experiments provide multi-benchmark evidence for the component interactions under heavy occlusion and long-tailed class distributions. The deployment-oriented YOLO-EMV model has also been integrated into customs security inspection equipment. Full article

This paper proposes a process-mapped quality control framework for structural steel fabrication as a proof of concept for integrating normative requirements into a unified operational workflow. The study addresses the fragmentation of quality requirements in steel fabrication, where individual standards are often applied through separate inspections rather than as part of a coherent process model. The framework links material control, production preparation, assembly, welding, and anticorrosion protection through clearly defined quality gates. The framework is constructed through a qualitative design-oriented approach using SIPOC analysis and a detailed flowchart. It translates normative requirements into a coherent workflow that provides a structured basis for traceability, preventive quality assurance, and future integration with Construction 4.0. The paper suggests that process mapping can serve as a useful method for organizing quality control in steel construction projects and for transforming standards into an operational framework. Full article

‘Wellness-at-sea’ is an integral part of the cruise holiday service bundle. Marketed within an image of experiential diversity within a safe, controlled environment, it features medical centers, spa facilities and restoration-focused itineraries. Nevertheless, epidemiologically speaking, cruise ships rank amongst the riskiest transmission environments in tourism. This perspective paper proposes a typology of wellness cruising along two axes (i.e., offer scope: themed voyages versus onboard amenities and promotion focus: borrowed credibility versus narrative ‘storyscaping’) and uses it to develop the ‘wellness-risk paradox’, a boundary case for health-tourism literature. Three testable propositions emerge: reputational risk exposure increases with the offer’s wellness centrality; narrative-based promotion is more susceptible than credibility-based promotion and a quadrant-adjusted health-assurance scheme is more effective than current inspection standards. The argument draws on the 2026 Andes hantavirus outbreak aboard the MV Hondius (three deaths among thirteen cases), the 2020 Diamond Princess outbreak, and a decade of official outbreak surveillance, reporting an increase in gastrointestinal outbreaks from 0.34 to 0.62 per million passengers between 2019 and 2025 (rate ratio 1.84, 95% CI 0.87–3.86), which reverses the pre-pandemic downward trend although annual counts remain too small for the change to reach conventional statistical significance. Sanitation-inspection compliance cannot reliably predict outbreaks, nor does it facilitate trust. Four domains of action are proposed: continuous air-quality monitoring, wastewater surveillance, behavioral risk modeling and jurisdictional coordination. Full article