Raspberry ketone is a valuable aromatic compound that occurs only at trace levels in natural sources. In this study, transgenic tomato fruits expressing benzalacetone synthase (BAS) from Rheum palmatum and raspberry ketone/zingerone synthase 1 (RZS1) from Rubus idaeus under the CaMV 35S promoter accumulated raspberry ketone (9.9-19.3 μg g-1 fresh weight) together with minor amounts of rhododenol (2.7-4.4 μg g-1 fresh weight), predominantly as glycosides during fruit ripening. LC-MS/MS analysis using an authentic standard identified raspberry ketone monoglucoside as a major glycosylated form present in transgenic fruits but absent in wild-type controls. Raspberry ketone glycosides were predominantly detected in the fruit peel, with negligible levels in the flesh, seeds, or leaves. Co-expression of the Arabidopsis thaliana R2R3-MYB transcription factor AtMYB12 further enhanced raspberry ketone production, reaching up to 88 μg g-1 fresh weight. These results demonstrate effective metabolic redirection of phenylpropanoid flux in an edible fruit tissue and establish tomato as a viable heterologous host for raspberry ketone production. This metabolic engineering strategy highlights the potential of edible fleshy fruits as sustainable plant-based platforms for the biosynthesis of high-value aromatic phenolics.
Metabolism-associated fatty liver disease (MAFLD) has emerged as a severe worldwide public health burden with insufficient available clinical therapeutic strategies, which underscores the urgent demand for safe, natural dietary interventions. Raspberry (Rubus idaeus L.), a typical food-medicine homologous fruit abundant in diverse bioactive components including anthocyanins, flavonoids and polysaccharides, possesses prominent nutritional and medicinal potential. In this study, raspberry aqueous extract (RE) was prepared to comprehensively investigate its ameliorative effects and underlying molecular mechanisms against MAFLD. MAFLD animal model was established in C57BL/6 mice via 12-week high-fat diet (HFD) feeding. From the 9th week, model mice were intragastrically administered with RE at doses of 1 g/kg/d and 2 g/kg/d for continuous intervention. Integrated multi-omics analyses including 16S rRNA microbial sequencing, serum/hepatic biochemical detection, histopathological examination, in vivo microbial colonization assay, and in vitro cellular and metabolomic experiments were performed to systematically clarify the regulatory mechanism. RE treatment markedly improved the core pathological phenotypes of MAFLD mice, and significantly mitigated hepatic steatosis and hepatocellular injury. 16S rRNA sequencing demonstrated that RE remodeled the gut microbial dysbiosis, specifically elevating the abundance of beneficial genus Ileibacterium and suppressing pathogenic microbial taxa. Meanwhile, RE strengthened intestinal mucosal barrier integrity by upregulating tight junction protein expression, and activated hepatic purine metabolic reprogramming to boost the levels of critical metabolites including inosine and ADP. Spearman correlation analysis verified the significantly positive correlation between Ileibacterium abundance and hepatic inosine content, and both factors were closely correlated with the remission of MAFLD pathological indicators. In vivo colonization experiments further validated that Ileibacterium intervention alone remarkably alleviated hepatic lipid deposition and liver damage in MAFLD mice. In vitro strain metabolomics confirmed that Ileibacterium could directly biosynthesize and secrete inosine extracellularly. Furthermore, in vitro AML12 hepatocyte experiments revealed that 100 μM inosine remarkably relieved palmitic acid-induced lipotoxicity via reducing intracellular lipid overload, reactive oxygen species (ROS) accumulation and mitochondrial dysfunction, alongside modulating the expression of lipid metabolism, inflammatory and autophagy-related genes. Collectively, our results elucidate that raspberry aqueous extract alleviates experimental MAFLD through the gut microbiota-purine metabolism-inosine regulatory axis, in which Ileibacterium and inosine act as the core synergistic mediators. This study provides solid preclinical experimental evidence for the development and application of raspberry as a promising functional food for the prevention and nutritional intervention of MAFLD.
Red raspberry (Rubus idaeus L.) cultivation is often constrained by drought stress, which reduces fruit yield and quality. Melatonin (MT) plays an important role in plant stress tolerance. This study examined how exogenous MT improves drought tolerance in red raspberry at the squaring stage. It focused on changes in photosynthetic performance, antioxidant capacity, leaves anatomical structure, gene expression, and hormone homeostasis. Drought stress significantly reduced photosynthetic performance, antioxidant enzyme activities, and hormone homeostasis in red raspberry, and altered leaves anatomical structure. Exogenous MT treatment, particularly at 150 μmol L⁻1, effectively alleviated these drought-induced effects. MT increased net photosynthetic rate and stomatal conductance, reduced oxidative damage, and promoted osmotic accumulation. At the anatomical level, MT-treated leaves showed more compact spongy tissue, increased palisade tissue and leaves thickness under drought stress. Transcriptomic and metabolomic analyses further revealed coordinated molecular changes under MT treatment. These changes mainly involved photosynthesis, carbohydrate and starch metabolism, plant hormone signaling, and amino acid metabolism. Genes related to photosynthetic function (PNSL3 and PNSB4), carbohydrate metabolism (glgc and PYG), antioxidant defense (SOD1 and sodC), and auxin signaling (AUX1 and SAUR) showed marked expression changes under MT treatment. Metabolite profiling also revealed changes in sugars, amino acids, phenolic compounds, and hormone-related compounds. Several transcription factor families, including bHLH, MYB-related, NAC, WRKY, and ERF, showed strong responses to MT. These changes were associated with transcriptional regulation and metabolic adjustment under drought stress. Exogenous MT significantly improved drought tolerance in red raspberry at the squaring stage, with 150 μmol L⁻1 showing the best overall performance. This protective effect was associated with improved photosynthetic performance, enhanced antioxidant capacity, and better maintenance of leaves structure. Integrated transcriptomic and metabolomic analyses further revealed changes in hormone-related and metabolic pathways. These findings indicate that MT enhanced drought tolerance in red raspberry through coordinated physiological, anatomical, and molecular responses.
Raspberry ketone synthase is the key enzyme for raspberry ketone biosynthesis, yet its activity and thermostability remain suboptimal for industrial use. By combining computational and deep-learning-guided design, we obtained a Rubus idaeus raspberry ketone synthase (RiRZS1) mutant M3 with 12.06-fold higher relative activity and 38.71-fold improved thermostability compared with the wild type. Kinetic analysis showed that M3 exhibited 14.89-fold higher kcat and 5.91-fold higher catalytic efficiency(kcat/Km), indicating markedly enhanced catalytic competence. Structural inspection and molecular dynamics simulations revealed that M3 induced only subtle conformational rearrangements but reduced global flexibility, accounting for its improved thermostability. Meanwhile, tunnel and substrate-pocket reshaping in M3 was associated with increased prereaction-state populations, which may contribute to enhanced catalytic performance. In whole-cell biotransformation, the raspberry ketone titer of M3 was 2.38-fold that of wild-type, and the corresponding space-time yield was 3.57-fold higher. This computational and artificial intelligence-assisted framework enabled a synergistic optimization of RiRZS1 activity and stability, and can serve as a transferable design framework for protein engineering in sustainable bio-manufacturing.
Foveolar-type gastric adenomas with a raspberry-like appearance in the stomach have recently been identified. However, their etiology and mechanisms of development and growth remain unknown. We observed a foveolar-type gastric adenoma with a raspberry-like appearance that underwent rapid morphological changes. The first esophagogastroduodenoscopy revealed multiple reddish polyps on the cardia and fundus of the stomach, without Helicobacter pylori infection. The second-largest polyp on the fundus had a red glove appearance and measured 5 mm in diameter. We diagnosed all of them as hyperplastic polyps on the basis of their shape. Given that the patient had received 10 mg of vonoprazan for a long time to treat reflux esophagitis, 20 mg of famotidine was prescribed instead. Eight months later, the largest polyp on the cardia had disappeared.However, a reddish raspberry-like lesion was observed at the site of a previously identified polyp on the fundus and was diagnosed as foveolar-type gastric adenoma by forceps biopsy. The pathological diagnosis was foveolar-type gastric adenoma with high-grade dysplasia as determined by tissue resection with endoscopic mucosal resection (EMR). This is a valuable case report of a lesion evolving from a hyperplastic-like polyp morphology to a raspberry-like lesion within eight months.
Foodborne viruses have caused large-scale outbreaks in fresh and frozen berries, such as strawberries, blueberries, and raspberries. Studies have previously evaluated the use of non-thermal technologies (including chemical sanitizers) to inactivate foodborne viruses on the surface of soft berries with varying results. In this study, we investigate a novel model system for utility in characterizing virus attachment and removal, as well as sanitizer inactivation on the surface of berries. A casting method using polydimethylsiloxane (PDMS) was used to create topomimetic artificial surfaces ("replicasts") of strawberries, blueberries, and raspberries to elucidate the impact of berry surface structures (topography) on virus adhesion and inactivation. A human norovirus surrogate, Tulane virus (TV), was inoculated (2 × 105 log10 PFU) on the surface of fresh and replicast berries, and viral recovery levels were compared. Overall, the type of berry impacted viral recovery, with 5.11 log10 PFU recovered from blueberries, 5.04 log10 PFU recovered from strawberries, and 5.66 log10 PFU recovered from raspberries. Surface inoculation of blueberry, strawberry, and raspberry replicasts with TV led to comparable recovery levels of virus as observed in fresh berries, with 4.80, 4.93, and 5.17 log10 PFU recovered, respectively, with a two-step recovery procedure. Sodium hypochlorite (200 ppm) led to less than a one log10 PFU reduction in viral titer in all three types of fresh berries. However, the same sanitizer treatment and two-step recovery procedure led to a 2.45 log10 PFU reduction in blueberry replicasts, 2.24 log10 PFU reduction in strawberry replicasts, and 1.59 log10 PFU reduction in raspberry replicasts. Sodium hypochlorite (50 ppm) and peracetic acid (20 or 80 ppm) had limited efficacy in inactivating viruses inoculated to fresh berry or berry replicast surfaces. However, viral reductions were greater in replicasts compared to fresh berries when observed. These data indicate that while berry replicasts may have utility in modeling surface adhesion of viruses to berries and in viral recovery studies, they are not suitable for studies of viral inactivation by chemical sanitizers.
The aim of the study was to modify polylactide with zinc oxide nanoparticles (ZnO), raspberry leaf extract (E), and a combined ZnO/extract system (EZnO) in order to prepare novel packaging materials via a solvent-free method, namely cast extrusion. Physicochemical properties: Morphology (GPC, SEM, FTIR), mechanical (tensile tests, puncture), barrier (WVTR, OTR, UV-Vis) and water contact angle for PLA-based films with two thickness ranges were investigated. Additionally, antimicrobial (antibacterial, antifungal and antiviral) tests were performed. GPC results revealed that the presence of the extract counteracted biopolyester degradation during hot melt processing. The best mechanical properties (TS ca. 50 MPa, EB ca. 18%) were obtained for PLA modified with raspberry leaf extract (PLA/E). EZnO addition led to the highest increase in oxygen (with 25%) and water vapor (up to ca. 28%) barrier properties. The material with EZnO addition was also found to be the only one to demonstrate antibacterial effectiveness, although the activity was insignificant. However, the incorporation of EZnO into the biopolymer matrix enhanced its antiviral properties, resulting in the complete inactivation of Φ6 bacteriophage particles used as a surrogate of SARS-CoV-2 virus.
This study demonstrated the effects of dietary supplementation with 1% raspberry (RO) or 1% strawberry (SO) seed oil from 5 to 12 weeks of age (n = 6/group) on folliculogenesis, hormonal parameters, the ovarian fatty acid profile, and the expression of related genes in juvenile rabbits. After slaughter, ovaries and blood were collected. Ovaries were used for histology, fatty acid profiling, and gene expression analysis, while plasma was used to measure progesterone (P4), testosterone (T), estradiol-17β (E2), follicle-stimulating hormone (FSH), and anti-Müllerian hormone (AMH) concentrations. Both RO and SO reduced the number of primary follicles (p = 0.04), whereas RO increased the number of antral follicles (p = 0.04) compared with the control. In both supplemented groups, FSH (p = 0.04 and p = 0.035) and AMH (p = 0.04) concentrations were higher. RO increased P4 and E2 (p = 0.03 and p = 0.013) concentrations, while SO only increased P4 (p = 0.02) levels. SO altered the ovarian fatty acid profile, increasing selected monounsaturated fatty acids and reducing polyunsaturated fatty acids, likely by increasing the expression of the converting enzyme, stearoyl-CoA desaturase 5 (p = 0.038). Overall, both oils influenced folliculogenesis through hormonal changes, and SO modified ovarian fatty acid composition, which may affect ovarian function in juvenile rabbits.
The tobacco constituent, N'-nitrosonornicotine (NNN) is a potential etiological agent in the development of oral cancer. NNN is metabolically activated to form DNA adducts that release 4-hydroxy-1-(3-pyridyl)-1-butanone (HPB) which was detected at higher levels in the oral cavity of smokers than non-smokers. Black raspberry (BRB) has been proposed as a preventive agent of oral cancer. In a phase 0 clinical trial, 40 active smokers (16 males, 24 females) consumed 5 g of BRB lozenges daily for 8 weeks. Each participant served as his/her own control. Buccal cells and urine were collected at baseline, during BRB administration, and washout period. In all smokers, BRB reduced, but not significantly, the levels of HPB-releasing DNA adducts (Baseline: 0.97 ± 0.51 and End-BRB: 0.61 ± 0.23 HPB/106 Guanine). A sex-specific response to BRB was observed: BRB had no effects in men, but in 16 of 24 women, BRB significantly reduced the levels of HPB. We further tested the hypothesis that detoxification of NNN via N-glucuronidation by BRB may, in part, account for the inhibition of HPB levels. The ratio of NNN-N-glucuronide to free NNN significantly increased (P = 0.028) from 1.29 ± 0.40 (baseline) to 2.77 ± 0.76 (End-BRB) in all 40 women as well as in female responders from 1.24 ± 0.60 (baseline) to 2.87 ± 1.00, but not in non-responders. The metabolic changes of NNN observed in our clinical trial indicate that BRB is a promising candidate for cancer interception and prevention of oral cancer in some, but not in all active smokers.
Accurate time synchronization is essential in distributed electrical signal monitoring, where phase coherence and event correlation depend on precise timing agreement between acquisition nodes. Conventional approaches often rely on a single synchronization source, typically internet-based Network Time Protocol (NTP) or GPS-disciplined clocks, which is impractical in isolated, offline, or cost-sensitive scenarios. This paper introduces an autonomous offline synchronization architecture for multi-node monitoring systems built on Raspberry Pi 5 (RPI5) platforms connected to a private Ethernet network. Instead of depending on one timing method, the system integrates several complementary mechanisms: battery-backed RTC persistence via the J5 interface, deterministic orchestration through systemd services, automated boot time recovery, chrony-managed NTP discipline, and Precision Time Protocol (PTP) hardware timestamping using PTP Hardware Clock (PHC). Synchronization performance is validated through continuous multi-day measurements of long-term stability, inter-node phase coherence, and short-term jitter. Controlled power-loss scenarios are also included to verify recovery behavior. The system maintains sub-microsecond alignment between nodes using only commodity hardware and no external time source. To further confirm inter-node timestamp alignment at the signal level, both hardware-based reference signal injection and software-based synchronized signal emulation are employed, providing ground-truth validation alongside scalable and reproducible evaluation. The results show that low-cost embedded hardware can support reliable, long-duration synchronization in fully offline installations.
High-fidelity manikin scenarios for contaminated airway management and upper gastrointestinal bleeding (UGIB) depend on blood or hematemesis simulants that reliably occlude optics and reproduce the visual gestalt of real blood under high-intensity video laryngoscopy and flexible endoscopy lighting. Commercially available simulated blood products are frequently optimized for reusability, stain resistance, and compatibility with procedural trainers, and can be cost-prohibitive for repeated high-volume "soiling" curricula. However, many simulated blood products do not report optical opacity or camera-obscuration performance for this specific use case. We performed a targeted technical review of blood and bleeding simulant strategies used in manikin-based training and adjacent materials science literature, and we describe a low-cost, reproducible hematemesis analog composed of three packets of a raspberry-flavored powdered beverage mix - Crystal Light Sugar-Free Raspberry Ice Drink Mix (made from maltodextrin, citrate salts, calcium phosphate, organic acids, and FD&C dyes) - dispersed with 20 g of unsweetened cocoa powder - Hershey's Cocoa - in 1 L of water using an immersion blender. In preliminary qualitative bench use, this particulate suspension formulation produced a dark red-brown appearance and visually reduced light transmission compared with the dye-only beverage mixture. Objective optical, rheologic, and comparative performance testing was not performed in this introductory report. Direct material cost was approximately $1.66/L using representative U.S. retail pricing, compared with $7-$29/L for selected commercial powders and premixes, although this comparison reflects direct consumable cost only and does not establish performance equivalence. Limitations include sedimentation and phase separation over time (necessitating re-homogenization) and potential staining of the manikin; a spot test on non-critical surfaces is recommended prior to implementation. Because the formulation contains organic particulates, it should not be introduced into patient-care endoscope working channels or equipment intended for later clinical use unless local reprocessing staff verify complete clearance for the specific device. This report is intended as a preliminary formulation and microcosting analysis; future work should quantify viscosity (including shear-thinning behavior), density, spectral absorbance/scattering, and material compatibility to support standardization across simulation programs.
The issue of parking management in university campuses has continued to face challenges owing to space constraints and the absence of real-time information. This problem is addressed by the proposed solution in this study, which is robust and intelligent and specifically designed for university campuses. This paper advances real-time campus parking through three innovations based on established You Only Look Once, Version 8 (YOLOv8) and Raspberry Pi tools: (1) campus-specific YOLOv8n fine-tuning (94.2% mean Average Precision (mAP), 450ms on Pi 4B); (2) adaptive Message Queuing Telemetry Transport (MQTT). Quality of Service (QoS) reducing 20% packet loss; and (3) S3-integrated forecasting producing 45% simulated efficiency gains. A user-friendly, web-based dashboard offers live parking updates to students, faculty, and visitors. This allows users to check space availability before arriving and reduces search time. The modular architecture supports decentralized deployment, with each Raspberry Pi independently managing a designated parking zone. The design is inherently scalable, enabling additional sensors and cameras to be added as needed to cover larger or more complex parking areas. To ensure privacy and reduce bandwidth use, live video and image access are restricted to management, maintaining data security and network efficiency. By combining edge computing, sensor fusion, and cloud services, the proposed solution enhances automation, improves user experience, and advances smart campus initiatives. This framework provides a scalable, adaptable model for modernizing parking infrastructure in educational institutions and beyond.
Drosophila suzukii Matsumura poses a significant threat to various soft-skinned fruit crops, causing substantial economic losses worldwide. Behavioral management using the repellent 2-pentylfuran has shown potential to manage this invasive pest. This 2-year study aimed to improve the effectiveness of 2-pentylfuran and identify cost-efficient deployment tools. In 2022, we evaluated automated canister puffers in isolated raspberry plantings and in semi-field cage trials using choice and no-choice assays. In 2023, we tested polyvinyl chloride spiral dispensers combined with below canopy insecticide applications in a large-scale raspberry planting, assessing D. suzukii infestations 72 h after 2-pentylfuran deployment. In addition, we monitored natural enemy and pollinator activities in both years using yellow sticky cards and visual counts, respectively. Treatments using 2-pentylfuran delivered by either puffers or spirals resulted in a small reduction in D. suzukii infestation, particularly in fall-bearing raspberries. However, combining 2-pentylfuran with insecticide in 2023 did not significantly improve efficacy. In semi-field trials, puffer density significantly reduced infestation in choice assays, while puffing frequency had no significant effect in either choice or no-choice assays. Importantly, 2-pentylfuran did not adversely affect the activity of beneficial insects and increased parasitism by the larval parasitoid Leptopilina japonica Novković & Kimura, potentially complementing biological control. Our findings suggest that further optimization of deployment strategies may facilitate the commercial integration of 2-pentylfuran into D. suzukii management programs.
Monitoring and identification of combustible gases are crucial tasks in fields such as industrial safety, environmental protection. In real-world scenarios, different types of combustible gases, such as H2, C2H4, and C2H2, often coexist, which makes recognition more difficult. To address this challenge, we propose a gas sensor based on a ZnO-Au-SnO2 heterostructure. The morphology and chemical state of the material were characterized using scanning electron microscopy and X-ray photoelectron spectroscopy, which revealed a uniform distribution of Au and binding-energy shifts indicative of electron transfer between ZnO and SnO2. Experimental results demonstrate that modifying Au and SnO2 significantly enhances the sensor's response to H2, C2H4, and C2H2. By applying a one-dimensional convolutional neural network (1DCNN) to a sensor array, efficient feature extraction of response signals and high-precision gas classification were achieved. The recognition accuracy for gas mixtures exceeds 96%. Furthermore, the lightweight 1DCNN model has been deployed on a Raspberry Pi edge computing platform, verifying its feasibility on small-sized, low-power devices. The system achieves millisecond-level inference and high-confidence predictions, indicating its potential for real-time on-site gas detection. This study synergistically optimizes the gas sensing system from both material and algorithm aspects, providing a feasible way to the intelligent recognition of multi-component combustible gases.
Understanding how the brain transforms sensory input and internal state into coordinated action requires behavioral paradigms that provide precise, multimodal measurements of movement and arousal while remaining compatible with neural recording techniques. Here we present a modular behavioral platform that enables stimulus-evoked locomotion in head-fixed mice using a transparent running wheel combined with air-stream stimulation. The design provides direct ventral access for imaging paw movements while simultaneously capturing body kinematics, facial motion, and eye-related signals from multiple camera views. The system integrates Arduino-based stimulus control, rotary encoder measurements, Raspberry Pi-based videography, and LED-based visual markers for event-level alignment across independently acquired data streams. Using a proof-of-principle dataset from well-trained animals, we show that brief air delivery reliably induces structured locomotion with reproducible trial timing. Optical-flow-based motion metrics and DeepLabCut pose estimation reveal robust, stimulus-locked increases in paw, limb, and facial movements during air-on epochs relative to air-off periods. LED-based event markers enable consistent event-level identification of air-on and air-off epochs across video streams despite differences in sampling rates. Together, these features provide a flexible framework for studying stimulus-driven locomotion and multi-view behavioral dynamics under head fixation, with straightforward compatibility for integration with neural imaging and electrophysiology recording approaches.
Accessible autonomous racing can engage undergraduate students emersed in artificial intelligence (AI) and robotics to receive an education through the excitement of maneuvering sharp corners and overtaking opponents. In particular, the construction and programming of miniature mobile racing robots can facilitate head-to-head racing competitions for use indoors within classrooms and hallways. Unfortunately, existing racing platforms remain inaccessible for overtaking maneuvers in such confined settings because they are physically large or expensive due to the computational cost of enabling such high performance. In an attempt to address these issues, we present Pocket Racer, an open-source, pocket-sized racing robot capable of head-to-head racing within indoor environments in education, i.e. university hallways or classrooms. We demonstrate head-to-head autonomous racing with our Pocket Racer platforms, enabling high speed overtaking upwards of 15 km/h. Designed to be easily assembled with off the shelf components and a low-cost edge device (Raspberry Pi Zero 2 W), our Pocket Racer platform is made accessible through an open source website and dataset detailing build instructions. By making a pocket-sized head-to-head autonomous racing platform accessible for undergraduate students, our work hopes to further hands on education in the age of physical AI.
Global crop movement has traditionally been viewed as a major driver of emerging plant diseases through the introduction of pathogens into naïve environments. Here we show that the reverse process, introducing crops into regions containing endemic pathogens already adapted to related native hosts, is an equally powerful but underrecognized mechanism of disease emergence. Using multilocus phylogeny, haplotype networks, SplitsTree analysis, and molecular clock dating of both fresh and century-old herbarium specimens, we reconstructed the global history of powdery mildews infecting strawberries and raspberries. We reveal that these fungi comprise ancient, geographically structured, host-specialized lineages rather than a single cosmopolitan species as previously assumed. North American lineages infecting strawberries (Podosphaera shepherdiae) and Eurasian lineages infecting strawberries (P. fragariae) trace their origins to native hosts, predating modern agriculture by millions of years. Raspberry-infecting lineages showed similar patterns of local endemism and host association. These findings demonstrate that emerging plant diseases can arise not only when pathogens move globally, but also when nonnative crops are introduced into landscapes containing long-established native pathogens. This work highlights the importance of taxonomic resolution and herbarium genomics for identifying the true origins of agricultural diseases and for understanding the evolutionary pathways that give rise to modern epidemics.
Structural cracks threaten the safety and long-term durability of civil infrastructure, yet manual inspection remains slow, subjective, and unreliable on complex surfaces. This study presents a training-centric strategy to improve a lightweight YOLOv11-N detector (2.6 M parameters) by transferring knowledge from a high-capacity YOLOv11-L teacher, without modifying the student architecture. Two semi-supervised mechanisms are investigated: pseudo-labeling and knowledge distillation. Using a crack dataset enlarged via systematic data augmentation, pseudo-labeling combines high-confidence teacher predictions with available ground-truth annotations through IoU and NMS filtering, while the distillation approach guides the student using both hard labels and teacher-derived soft signals to strengthen localization behavior. Experimental results show that both strategies enhance the baseline student model, with pseudo-labeling providing more stable training dynamics and stronger overall gains, whereas distillation primarily improves convergence behavior and sample efficiency. Ablation analyses highlight that the benefit of pseudo-labeling is data-dependent and requires a minimum pseudo-label density to achieve consistent improvements. Finally, edge-device (Raspberry 3B + and Nvidia Jetson Nano Kit) benchmarks validate that the resulting lightweight detector is suitable for deployment on resource-constrained platforms, enabling practical UAV- and mobile-oriented crack inspection across diverse edge computing tiers.
Background/Objectives: Brain health monitoring is increasingly essential as modern cognitive load, stress, and lifestyle pressures contribute to widespread neural instability. The paper presents BrainTwin, a next-generation cognitive digital twin, as a patient-specific, constantly updating computer model that combines state-of-the-art MRI analytics for neuro-oncological assessment related to clinical study and management of tumors affecting the central nervous system (including their detection, progression, and monitoring) with real-time EEG-based brain health intelligence. Methods: Structural analysis is driven by an Enhanced Vision Transformer (ViT++), which improves spatial representation and boundary localization, achieving more accurate tumor prediction than conventional models. The extracted tumor volume forms the baseline for short-horizon tumor progression modeling. Parallel to MRI analysis, continuous EEG signals are captured through an in-house wearable skullcap, preprocessed using Edge AI on a Hailo Toolkit-enabled Raspberry Pi 5 for low-latency denoising and secure cloud transmission. Pre-processed EEG packets are authenticated at the fog layer, ensuring secure and reliable cloud transfer, enabling significant load reduction in the edge and cloud nodes. In the digital twin, EEG characteristics offer real-time functional monitoring through dynamic brainwave analysis, while a BiLSTM classifier distinguishes relaxed, stress, and fatigue states, which are probabilistically inferred cognitive conditions derived from EEG spectral patterns. Unlike static MRI imaging, EEG provides real-time brain health monitoring. The BrainTwin performs EEG-MRI fusion, correlating functional EEG metrics with ViT++ structural embeddings to produce a single risk score that can be interpreted by clinicians to determine brain vulnerability to future diseases. Explainable artificial intelligence (XAI) provides clinical interpretability through gradient-weighted class activation mapping (Grad-CAM) heatmaps, which are used to interpret ViT++ decisions and are visualized on a 3D interactive brain model to allow more in-depth inspection of spatial details. Results: The evaluation metrics demonstrate a BiLSTM macro-F1 of 0.94 (Precision/Recall/F1: Relaxed 0.96, Stress 0.93, Fatigue 0.92) and a ViT++ MRI accuracy of 96%, outperforming baseline architectures. Conclusions: These results demonstrate BrainTwin's reliability, interpretability, and clinical utility as an integrated digital companion for tumor assessment and real-time functional brain monitoring.
This paper presents a novel edge-deployable assistive system for visually impaired individuals, powered by vision-language models (VLMs). Existing cloud-based solutions for image captioning suffer from latency, dependency on internet connectivity, and overly simplistic scene descriptions that fail to convey the rich contextual information needed for meaningful real-world navigation. To address these challenges, we propose a self-contained, wearable system that performs real-time scene interpretation on-device without cloud reliance. The system integrates a quantized version of the LLaVA-NeXT-13B VLM with speech recognition (Whisper) and speech synthesis (PIPER), running on an NVIDIA Jetson Orin NX and Raspberry Pi-based input module. Our framework emphasizes intuitive user interaction through a button-based interface and Bluetooth audio output, minimizing cognitive load. We validate the quantization approach through large-scale benchmarks, such as VizWiz-VQA and VQAv2, demonstrating minimal accuracy degradation (2.6% and 0.9%, respectively) compared to the original model. User evaluations involving 28 participants compared the system to a baseline image captioning model. Objective results demonstrated a 25% increase in image identification accuracy. The system achieved high usability scores and maintained practical latency (4-5 seconds per query), supporting real-world feasibility. This work advances the development of scalable, interpretable, and accessible AI-driven assistive technologies, enabling greater independence and interaction for the visually impaired.