Obstructed mixed-type total anomalous pulmonary venous connection (TAPVC) is a rare life-threatening congenital heart defect requiring urgent, high-risk surgery. Postoperative care is complex and fraught with complications. Enhanced recovery after surgery (ERAS) is a multimodal, evidence-based approach to optimize perioperative care, yet its application in neonatal cardiac surgery is nascent. A 12-day-old male infant presented with lethargy, poor feeding, and severe respiratory distress (SpO 2 80%-85%), requiring immediate intubation. Examination revealed diminished mental responsiveness and signs of poor perfusion. Obstructed mixed-typeTAPVC with an atrial septal defect, a ventricular septal defect, a patent ductus arteriosus, and severe pulmonary hypertension. The infant underwent surgical correction. Postoperatively, a comprehensive, nursing-led ERAS protocol was implemented. Core interventions encompassed goal-directed fluid therapy, proactive respiratory support to prevent pulmonary hypertensive crises, standardized sedation, early complication surveillance (thrombosis, capillary leak syndrome), and structured early rehabilitation. The infant was successfully weaned from mechanical ventilation on postoperative Day 5, achieved full oral feeding by Day 13, and was discharged on Day 18. This recovery timeline compares favorably to published data for similar complex cases. At the 3-month follow-up, he demonstrated age-appropriate growth and development with no significant cardiac sequelae. A structured, nursing-driven ERAS protocol is a valuable clinical tool to enhance outcomes, accelerate recovery, and reduce complications in high-risk neonatal cardiac surgery. This case affirms the pivotal role of neonatal nurses in perioperative care.

Zinner syndrome is a rare congenital defect of the male reproductive organs, which is traditionally characterised by the presence of the triad of unilateral renal agenesis, ipsilateral seminal vesicle cyst, and obstruction of ejaculatory ducts. It is a result of malformed development of the mesonephric (Wolffian) duct during embryogenesis, commonly asymptomatic, and is discovered incidentally by imaging. Mass effect can cause symptomatic patients to present with lower urinary tract symptoms, lower abdominal or pelvic pain, reproductive dysfunction, and bowel symptoms, uncommonly. Magnetic Resonance Imaging (MRI) is the diagnostic gold standard, Computed Tomography (CT), Ultrasound (US) and emerging modalities can help in early detection. Individualised management includes conservative observation and minimal invasive, robotic-assisted or endoscopic intervention with new therapies emerging to help relieve the symptoms and preserve fertility. Genetic research, such as whole exome sequencing, has proposed a role of genetic factors such as GATA3 and PAX8, among others, in the pathogenesis. The present article is a review of clinical presentation, imaging, pathogenesis, and treatment.

We report an uncommon presentation and conservative management of Aplasia Cutis Congenita (ACC), a rare congenital defect of the skin in a newborn. The skin defects was located on the abdominal region, irregularly shaped, symmetrical, each measuring about 5 x 3 cm. There was no discharge, bleeding, bullae or anomalies on clinical examination. The child was managed conservatively with antibiotics and wound dressing. The wound granulated well, there were no local or systemic complications. By the 8th day of admission, defect had almost completely healed leaving a hypo-pigmented scar. At 2 weeks on follow-up the scar had become hyper-pigmented.

Background: Supravalvular aortic stenosis (SVAS) is a rare congenital heart defect first described by Chevers in 1842. It accounts for approximately 8–14% of all congenital aortic stenosis cases and represents 0.05% of all congenital heart diseases. Surgical repair remains an effective primary treatment for this condition. Case Presentation: This report describes a clinical case of SVAS successfully diagnosed and treated at Viet Duc University Hospital. Results: A 27-year-old male patient was diagnosed with SVAS and underwent successful surgical reconstruction of the aortic root using the modified three-patch technique (Brom technique). Postoperative computed tomography (CT) scan at the 3-month follow-up showed favorable anatomical and hemodynamic outcomes. Conclusion: Supravalvular aortic stenosis (SVAS) is a rare congenital condition associated with significant risks and potential complications. Aortic root reconstruction has proven to be an effective surgical approach, providing favorable long-term outcomes for patients.

Sinus venosus atrial septal defect is a rare congenital heart defect that is frequently associated with partial anomalous pulmonary venous return (PAPVR). This defect is categorized into 2 types: superior vena cava (SVC) type and inferior vena cava type, with the SVC type being more common. The condition results in abnormal left-to-right shunting of blood, leading to right heart dilation and increased pulmonary blood flow. A comprehensive literature search was performed using PubMed, MEDLINE Ovid, and Embase, focusing on studies involving patients with sinus venosus atrial septal defect, with or without PAPVR, published in English. Data were collected on patient demographics, the type of sinus venosus atrial septal defect, comorbidities, diagnostic imaging methods, and treatment outcomes. The search aimed to gather information on the clinical presentations, diagnostic challenges, and treatment options for this congenital defect. Sinus venosus atrial septal defect accounts for 5–10% of all atrial septal defects. Most cases involve the SVC type, which is often associated with PAPVR. Patients commonly present with symptoms such as dyspnea, fatigue, palpitations, and exercise intolerance. Comorbidities such as Lutembacher syndrome, cor triatriatum dexter, and pulmonary hypertension were observed. Diagnostic methods included transthoracic echocardiography, transesophageal echocardiography, cardiac magnetic resonance imaging, and computed tomography angiography, with transesophageal echocardiography being the most sensitive. Sinus venosus atrial septal defect is a clinically significant condition that can lead to serious complications like atrial arrhythmias and pulmonary hypertension. Early diagnosis using advanced imaging techniques and surgery is essential for preventing long-term morbidity.

As the cornerstone of contemporary electronics, the quality of electronic substrates—including Printed Circuit Boards (PCBs) and Ceramic Package Substrates (CPSs)—is intrinsic to product reliability. However, automated inspection is currently impeded by two persistent obstacles: the drastic multi-scale variation in defects and the acute class imbalance within defect datasets. Conventional deep learning approaches often fail to reconcile these challenges simultaneously, leading to suboptimal recognition of rare defect categories. To bridge this gap, we propose Multi-scale Partial Vision Transformer—Simple, Parameter-free Attention Module (MulPViT-SimAM), a robust framework designed for class-imbalanced electronic substrate defect detection. Our method features a novel multi-scale backbone (MulPViT) that synergizes partial convolutions with hierarchical attention mechanisms, facilitating the efficient extraction of both fine-grained local textures and global contextual dependencies. Additionally, we embed the Simple, Parameter-free Attention Module (SimAM) into the feature fusion stage to adaptively highlight defect-specific features while dampening background noise. To further mitigate data imbalance, we utilize the Equalized Focal Loss (EFL) function, which employs a category-specific modulating factor to dynamically equilibrate the learning focus across different classes. Comprehensive benchmarking reveals state-of-the-art performance, achieving mAP@0.5 scores of 95.7% on the standard PKU-MARKET-PCB dataset and 54.2% on the highly challenging CPS2D-AD dataset. Significantly, our approach effectively mitigates class imbalance, narrowing the performance deviation of rare categories to just 4.3% on the PKU-Market-PCB dataset and 1.4% on the CPS2D-AD dataset, compared to 11.8% and 7.5% in baseline models. These findings position MulPViT-SimAM as a viable and efficient solution for industrial quality control.

Background: Iniencephaly is an extremely rare and fatal neural tube defect characterized by severe malformations of the craniovertebral junction, including absence or defect of the occipital bone, marked retroflexion of the head, and abnormalities of the cervical and upper thoracic spine. It is frequently associated with multiple congenital anomalies and is incompatible with postnatal survival. Early prenatal diagnosis is essential for appropriate counseling and management. Case Presentation: A 14-week gestation fetus of a 28-year-old primigravida was antenatally diagnosed with Iniencephaly. Ultrasonography revealed persistent hyperextension of the head, absent cervical curvature, an open craniospinal neural tube defect, and features suggestive of omphalocele. Following medical termination, fetal autopsy showed absence of skull bones, extreme retroflexion of the head with absent neck, and a continuous neural tube defect extending from the occipital to lumbar region (craniorachischisis). Associated anomalies included omphalocele, bilateral ventriculomegaly, absence of the stomach, and Potter’s facies. Infantogram confirmed absent cranial vault ossification with relatively preserved axial skeleton. Conclusion: The case demonstrates the classical pathological features of iniencephaly with multiple congenital anomalies and its uniformly fatal outcome. Early antenatal detection is crucial for diagnosis, counseling, and prevention through periconceptional folic acid supplementation.

Congenitally corrected transposition of the great arteries is a rare congenital heart defect that may remain undiagnosed well into adulthood. We present the case of a 71-year-old male with dextrocardia and exertional dyspnoea, initially diagnosed with severe pulmonary hypertension. Further evaluation revealed an underlying congenitally corrected transposition of the great arteries. This case highlights the importance of considering CHD in older adults with unusual cardiac findings.

• Addresses signal processing limits from class-imbalanced DED molten pool signals. • Physics-guided CVAE-GAN synthesizes high-fidelity visual signals. • Thermodynamic losses enforce spatial and spectral signal consistency. • Augmented data improves decision boundaries for rare defect detection. • Validated on a real-world aerospace component with complex geometry. Vision-based signal processing is critical in metal additive manufacturing (AM), particularly within Directed Energy Deposition (DED), for capturing the rapid thermo-fluid dynamics essential for detecting process instabilities and ensuring structural integrity. While DED operates as an open, inherently non-stationary system and molten pool visual signals carry rich spatiotemporal information, effective processing is fundamentally constrained by the scarcity of anomalous events relative to stable deposition. This data sparsity leads to imbalanced training sets that fail to represent the full spectrum of transient process dynamics, causing classifiers to misclassify rare but critical failure modes and degrading overall monitoring performance. Crucially, conventional data augmentation techniques, being agnostic to the underlying physics, often introduce signal artefacts that violate governing thermodynamic constraints, thereby degrading the physical validity of the training data. To bridge the gap between data-driven feature learning and physical principles, this study proposes a physics-guided generative data augmentation framework based on a Conditional Variational Autoencoder–Generative Adversarial Network (CVAE-GAN). The framework explicitly learns the intrinsic physical manifold of the process, incorporating thermodynamically motivated loss functions to ensure the synthesis of high-fidelity, physically coherent molten pool visual signals. Validation through both spectral analysis and downstream classification confirms that the augmented signals preserve essential morphological and textural properties while significantly improving the decision boundaries of sequence classifiers. A case study on a complex aerospace component demonstrates the framework’s practical utility in enhancing robust in-situ monitoring under varying geometric conditions. By addressing the sparse representation of critical dynamics through a physics-aware methodology, this work offers a reliable pathway for advanced signal processing in data-scarce engineering systems.

Abstract Background Congenitally corrected transposition of the great arteries (ccTGA) is a rare congenital heart defect characterized by atrioventricular and ventriculoarterial discordance and frequently associated with heterogeneous anatomical anomalies. This variability makes surgical indication and strategy particularly complex, requiring accurate preoperative anatomical and functional assessment. Case series We report four pediatric patients with complex ccTGA who underwent anatomical repair at our institution between July 2022 and March 2024. Associated findings included dextrocardia, situs inversus, ventricular septal defects, pulmonary outflow tract obstruction, and atrioventricular valve abnormalities. Imaging and methods All patients underwent stepwise multimodal imaging. Transthoracic echocardiography was complemented by cardiovascular magnetic resonance (CMR) to assess ventricular volumes and function, septal defects, outflow tract relationships, systemic and pulmonary venous return. CMR datasets were processed to generate patient-specific virtual and 3D-printed models through dedicated segmentation, preserving real anatomical dimensions. These models were used for surgical simulation and multidisciplinary planning. Findings Integration of CMR and 3D modeling directly influenced operative strategy in all cases. Three-dimensional reconstructions refined ventricular septal defect tunneling toward the aorta, optimized ventriculotomy site selection, clarified atrial switch configuration, and anticipated the need for pulmonary root translocation. Anatomical repair was successfully achieved in all four patients, with no early mortality and preserved biventricular function at short follow-up. Conclusions Structured integration of CMR and patient-specific 3D modeling enhances spatial understanding and supports individualized surgical planning in complex ccTGA. Incorporating advanced imaging into routine clinical workflow represents a valuable strategy to improve precision and decision-making in anatomically challenging congenital heart disease.

Patient: Female, 44-year-oldFinal Diagnosis: Severely calcified truncal valve with moderate regurgitation • uncorrected truncus arteriosus type 1 • ventricular septal defect (VSD)Symptoms: Aphasia • neurological deficits • right-sided weakness, and vertigoClinical Procedure: —Specialty: CardiologyObjective: Congenital defects/diseasesBackgroundIn adults, non-corrected truncus arteriosus type 1 with Eisenmenger syndrome is a rare congenital heart defect characterized by a single large artery from the heart that directs blood flow to the lungs, leading to irreversible pulmonary hypertension and a reversal of blood flow. This report describes the case of a 44-year-old woman with congenital Eisenmenger syndrome and unrepaired truncus arteriosus type 1 presenting with recurrent thromboembolic cerebral infarcts.Case ReportA 44-year-old woman presented with a 2-month history of recurrent neurological deficits, including aphasia, right-sided weakness, and vertigo. Magnetic resonance imaging (MRI) of the brain confirmed acute embolic infarcts, while computed tomography (CT) angiography ruled out large-vessel occlusion. Cardiac evaluation revealed a 15-mm perimembranous ventricular septal defect with a bidirectional shunt and a severely calcified truncal valve with moderate regurgitation. Imaging confirmed truncus arteriosus type 1, with the pulmonary arteries originating from the ascending aorta. The right ventricular systolic pressure was estimated at 114 mmHg, and baseline oxygen saturation (SpO2) was 70–80%. Due to irreversible pulmonary vascular disease, she was not a candidate for surgical repair. A conservative management strategy was adopted, including dual antiplatelet therapy, sildenafil, macitentan, bisoprolol, and oxygen therapy. At the 6-month follow-up, the patient remained neurologically stable with her SpO2 improved to 85–88%.ConclusionsThis case highlights the exceptional natural course of uncorrected truncus arteriosus in adulthood. It underscores that for patients with Eisenmenger physiology where surgical repair is not feasible, a multidisciplinary approach focusing on tailored antithrombotic and pulmonary vasodilator therapies is essential for stroke prevention and clinical stability.

Aortopulmonary window (APW) is a rare congenital heart defect, accounting for only 0.2-0.6% of all congenital cardiac anomalies, and is usually identified and surgically corrected within the first year of life to prevent irreversible pulmonary hypertension (PH). Once pulmonary vascular resistance (PVR) becomes fixed, surgical repair is generally considered contraindicated. However, emerging evidence suggests that pulmonary vascular reactivity may persist in select adolescents, opening a potential therapeutic window beyond infancy. We report a 15-year-old male with a large APW diagnosed in infancy but lost to follow-up, who presented in adolescence with exertional dyspnea. Initial cardiac catheterization revealed a high indexed PVR (10.89 WU·m²) and absent oxygen vasoreactivity, indicating inoperability. Despite this, preserved biventricular function and lack of cyanosis prompted a trial of targeted medical therapy with sildenafil, spironolactone, lisinopril, and digoxin. After 12 months, repeat catheterization showed dramatic haemodynamic improvement: baseline PVRi decreased to 8.84 WU·m² and fell further to 0.76 WU·m² after 100% oxygen, with a reduction in PVR/SVR ratio from 0.58 to 0.04. The calculated Qp/Qs increased to 16.77, although this extreme value was considered likely overestimated because systemic and pulmonary arterial oxygen saturations were almost identical. Definitive surgical repair was undertaken with excellent early results. On follow-up, the patient remained asymptomatic with preserved biventricular function, low estimated pulmonary artery pressures and no residual APW on echocardiography. This case suggests that operability in APW may not be irrevocably lost beyond infancy in carefully selected adolescents. It underscores the importance of individualized assessment, medical preconditioning and serial haemodynamic reassessment to unmask latent pulmonary vasoreactivity. Larger series are needed to define which late-presenting patients with APW and PH may safely benefit from definitive surgical repair.

Rationale:Hysterosalpingo-contrast sonography (HyCoSy) is a commonly used technique for assessing fallopian tube patency. A well-recognized complication of this procedure is venous intravasation (VI), which refers to the unintended entry of contrast medium into the uterine and parametrial venous systems. This highly variable complication often leads to misdiagnosis, which in turn compromises the subsequent management and treatment of affected patients.Patient concerns:A 32-year-old female who underwent a 4-dimensional HyCoSy for secondary infertility initially appeared to have a rare congenital defect, with images revealing three distinct tube-like structures resembling fallopian tubes, one on the right side of the uterus and the other two on the left side, which is suspicious for fallopian tube duplication malformation.Diagnoses:Real-time 2-dimensional imaging and dynamic cine-loop recordings revealed that the contrast agent originated from myometrial vessels rather than the tubal ostia. This confirmed a diagnosis of extensive VI mimicking fallopian tube duplication malformation, coupled with bilateral tubal obstruction.Interventions:The diagnosis was substantiated by the absence of peritoneal contrast spillage and the observation of delayed ovarian stromal enhancement, a sign of systemic recirculation. The findings were further verified via hysteroscopy and an intra-operative chromopertubation test. Following this, the patient was referred for in vitro fertilization-embryo transfer.Outcomes:After receiving appropriate fertility treatment based on the corrected diagnosis, the patient successfully achieved pregnancy via in vitro fertilization-embryo transfer.Lessons:Significant VI can manifest as organized “pseudo-tubular” structures that convincingly mimic congenital malformations. To ensure diagnostic precision, clinicians must track the anatomical origin of contrast egress and look for indirect signs of intravasation, such as delayed ovarian enhancement, to avoid unnecessary invasive procedures and guide correct clinical decision-making.

Introduction: A coronary sinus atrial septal defect, also known as an unroofed coronary sinus (UCS), is an exceedingly rare congenital anomaly with rates of less than 1% of all atrial septal defects. This defect often presents with left to right shunting and may be misdiagnosed as idiopathic pulmonary arterial hypertension (PAH) leading to right heart failure. Description: A 75 year-old female with diagnosed with idiopathic PAH presented for evaluation of dyspnea, lower extremity edema, and signs of right heart failure which was confirmed on initial workup. Subsequent transthoracic echocardiography demonstrated a severely dilated right atrium, right ventricle (RV), and left atrium with severely reduced RV systolic function and preserved left ventricular function. CT was reviewed again which revealed a possible UCS and this was confirmed on cardiac MRI as a 1.8cm wide coronary sinus atrial septal defect with no evidence of persistent left superior vena cava. This is known as a type II UCS. She underwent a right heart catheterization which showed left to right shunting with an oxygen step-up from 68% to 90.6% in the RA and a calculated shunt ratio of 2.2 and Qp/Qs of 2.4. The patient’s symptoms improved with diuresis, and she was referred for surgical correction. Discussion: This case highlights the challenge posed by rare congenital defects in adults presenting with unexplained PAH or acute right heart failure. The delayed recognition and absence of a persistent left superior vena cava further illustrates the need for multimodal imaging and reasonable clinical suspicion. Early identification and repair may improve symptoms and percent long term complications such as paradoxical emboli and progressive RV failure.

Introduction: Hypoplastic left heart syndrome (HLHS) is a rare congenital cardiac defect in which systemic circulation is dependent on right ventricular output. The Fontan procedure creates a pathway for blood to circulate directly into the lung, which ultimately results in persistent shunt physiology. The complex hemodynamic physiology that arises is especially challenging in critically ill patients, where adequate oxygenation is of utmost importance. Inhaled pulmonary vasodilators such as epoprostenol (EPO) can selectively reduce pulmonary vascular resistance (PVR), improve ventilation-perfusion matching, and optimize systemic oxygen delivery. Description: 32-year-old male with HLHS status post-Fontan presented with progressive dyspnea following a week of nausea and vomiting. He transitioned from an oxymizer to high-flow nasal cannula (40 L/min, FiO2 1.0), then to BiPAP (12/5, FiO2 1.0), but remained hypoxic. ABG on BiPAP showed PaO2 50 mmHg. He was intubated for worsening respiratory failure, but oxygenation declined with PEEP of 8 cm H2O (PaO2 54 mmHg on FiO2 0.8, SpO2 92.5%), likely due to impaired pulmonary blood flow from elevated intrathoracic pressure. Inhaled epoprostenol was initiated, resulting in marked improvement in oxygenation. The patient was weaned to HFNC with epoprostenol, then transitioned as low as HFNC 30L/.45 FIO2 and transferred to a congenital heart center for further management. Discussion: In Fontan patients, positive pressure ventilation can worsen oxygenation by impeding passive pulmonary blood flow and increasing right-to-left shunting. Inhaled epoprostenol improves V/Q matching by reducing PVR and enhancing pulmonary flow without systemic vasodilation. This case demonstrates the utility of epoprostenol as a rescue strategy for refractory hypoxemia in adults with complex congenital heart disease. Recognition of unique physiologic interactions in this population is crucial for individualized management in the ICU.

Wafer map defect classification plays a critical role in yield monitoring and root-cause analysis in semiconductor manufacturing. Although recent convolutional neural network (CNN)-based approaches have achieved high classification accuracy, most existing models are evaluated primarily on clean datasets and remain vulnerable to unseen perturbations and representation-level variability at test time. In this paper, we propose a hybrid CNN-echo state network (ESN) architecture that integrates spatial feature extraction with sequential aggregation to enhance robustness under input perturbations. The CNN backbone extracts two-dimensional feature maps, which are converted into ordered sequences using a multidirectional scanline strategy and processed by an ESN reservoir. The resulting sequential representations are combined with CNN features through a class-specific adaptive fusion mechanism. Using the defect-only eight-class version of the WM-811K dataset, we systematically evaluate robustness under multiple perturbation scenarios, with particular focus on the clean train/noisy test (CT-NT) setting. To ensure a controlled robustness evaluation aligned with the binary nature of wafer map data, we introduce binary-consistent die-flip perturbations and additionally employ additive Gaussian perturbations as a representation-level stress test. Under clean-data conditions, the proposed model showed a 0.61 pp improvement in test accuracy compared to the ResNet34-based CNN, with notably larger gains for rare classes and defect types exhibiting strong structural patterns. In the clean train/noisy test scenario, where the model was trained on clean wafer map data and evaluated under controlled test-time perturbations, the accuracy of the CNN baseline dropped to 77.59% at σ = 0.10, whereas the proposed hybrid model maintained an accuracy of 87.30%, resulting in an absolute improvement of 9.71 pp. Per-class analysis reveals that the robustness gain is class-dependent, with pronounced improvements for defect types exhibiting clear and repetitive structural patterns, such as Loc and Edge-Ring. Further mechanistic analysis demonstrates that the robustness improvement arises from enhanced representation stability and bounded reservoir dynamics, rather than from changes in CNN feature extraction or training regularization. These results demonstrate that the proposed CNN-ESN hybrid architecture provides meaningful advantages in terms of robustness under noisy evaluation conditions without requiring noise-aware training or prior knowledge of perturbation characteristics.

Lumbar hernias are rare abdominal wall defects, either primary or secondary, with symptoms like flank masses or bowel obstruction. Though uncommon, incarceration occurs in about 9% of cases, necessitating early surgical intervention. While both open and laparoscopic repairs are viable, laparoscopic methods, especially the totally retroperitoneal endoscopic approach, have shown promising outcomes despite the technical challenges associated with this technique. Over the past 5 years, we employed this technique in various lumbar hernias, with favorable recovery results, highlighting its clinical potential. Patients with lumbar hernias, each located within the superior lumbar triangle, were prospectively enrolled in this study. All hernias were repaired using the totally extraperitoneal (TEP) technique, which allows for complete extraperitoneal access. Detailed descriptions of the surgical technique are provided, and perioperative outcomes were systematically recorded and analyzed. Seventeen patients underwent TEP repair for lumbar hernias, including one with bilateral hernias. The cohort had a mean age of 62 years and a mean BMI of 26.1 kg/m2. Hernia defects ranged from 2 to 6 cm, with equal proportions of primary and secondary types. The median hospital stay was 3 days, and the median postoperative pain score was 3. Over 6-48 months of follow-up, no recurrences or serious complications were observed. This study demonstrated that the TEP approach is a feasible, safe, and technically straightforward method for repairing upper lumbar hernias. Given the rarity of lumbar hernias, larger studies with extended follow-up are warranted to confirm the long-term efficacy of this technique.

Uhl's anomaly is an extremely rare congenital heart defect, with approximately 100 cases reported since its first description in 1952 by Dr Henry Uhl. Characterized by partial or complete absence of the right ventricular myocardium, it leads to progressive right ventricle dilation and right heart failure. Due to its rarity, there is no consensus regarding its surgical or medical management. We present a case of an 18-year-old male patient with Uhl's anomaly diagnosed in early infancy, managed by palliative Glenn shunt. This report provides insight into management of this rare condition.

ABSTRACT Quality assessment in additive manufacturing (AM) relies on defect detection models, whose performance is limited by the lack of defect data and the high imbalance between classes. To address this, this paper introduces a physics-informed multimodal generative adversarial network (GAN) framework for synthetic defect data augmentation. It combines three innovations including physics-constrained models with thermodynamic consistency between visual and thermal representations, cross-modal attention mechanisms with realistic correlations between data features and adaptive augmentation strategies with dynamically computed synthetic-to-real ratios depending on model uncertainty and validation accuracy. The framework, when evaluated on a real-world AM process monitoring dataset of 1500 high-resolution layer images, achieved 89.2% defect detection accuracy, a 12.8% improvement over no augmentation and a 6.7% improvement over state-of-the-art baseline models. On average, recall for rare defect detection increased by 31.6%, particularly for the most severe cases, where the training prevalence was less than 5%. The documentation openly describes an initial training failure, with the discriminator achieving a fake accuracy of 1.00. The paper then details the systematic corrections applied: reducing the discriminator learning rate by 5x, automatically optimising the batch size and adding a regularisation method. This transparent account of failure and resolution makes the documentation highly valuable to practitioners.

The structural integrity of aero-engine blades is paramount to aviation safety, requiring rigorous inspection protocols to identify surface and sub-surface defects. While deep learning has revolutionized industrial defect detection, its efficacy is severely constrained in aerospace applications by the paucity of defect samples, a condition known as the few-shot learning problem. Standard convolutional neural networks struggle to generalize from limited data, leading to overfitting and poor recognition of rare defect classes. This paper proposes a novel framework that integrates Generative Feature Enhancement with Metric Learning to address the scarcity of labeled defect data in aero-engine blades. Unlike traditional data augmentation techniques that operate in the image space, our approach synthesizes diverse training samples in the high-dimensional feature space using a modified generative adversarial network. Concurrently, a metric learning module optimizes the embedding space to ensure that defects of the same class are clustered tightly while maintaining significant separation from non-defect samples and other defect categories. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art few-shot learning algorithms in identifying micro-cracks, erosion, and ablation on blade surfaces. The dual-branch architecture ensures robust generalization capabilities even when training examples are reduced to single digits per category.