Search Results (624)

Background/Objectives: This proof-of-concept study evaluated whether optical coherence tomography angiography (OCTA) can non-invasively capture micro-vascular alterations in non-melanoma skin cancer (NMSC) lesions during and after superficial orthovoltage radiotherapy (RT) using radiomics and vascular features analysis. Methods: Eight patients (13 NMSC lesions) received 36–50 Gy in 6–20 fractions. High-resolution swept-source OCTA volumes (1.1 × 10 × 10 mm³) were acquired from each lesion at three time points: pre-RT, immediately post-RT, and three months post-RT. Additionally, healthy skin baseline was scanned. After artifact suppression and region-of-interest cropping, (i) first-order and texture radiomics and (ii) skeleton-based vascular features were extracted. Selected features after LASSO (least absolute shrinkage and selection operator) were explored with principal-component analysis. An XGBoost model was trained to classify time points with 100 bootstrap out-of-bag validations. Kruskal–Wallis tests with Benjamini–Hochberg correction assessed longitudinal changes in the 20 most influential features. Results: Sixty-one OCTA volumes were analyzable. LASSO retained 47 of 103 features. The first two principal components explained 63% of the variance, revealing a visible drift of lesions from pre- to three-month post-RT clusters. XGBoost achieved a macro-averaged AUC of 0.68 ± 0.07. Six features (3 texture, 2 first order, 1 vascular) changed significantly across time points (adjusted p < 0.05), indicating dose-dependent reductions in signal heterogeneity and micro-vascular complexity as early as treatment completion, which deepened by three months. Conclusions: OCTA-derived radiomic and vascular signatures tracked RT-induced micro-vascular remodeling in NMSC. The approach is entirely non-invasive, label-free, and feasible at the point of care. As an exploratory proof-of-concept, this study helps to refine scanning and analysis protocols and generates knowledge to support future integration of OCTA into adaptive skin-cancer radiotherapy workflows. Full article

Multiple-mode immunoassays have the advantages of self-correction, self-validation, and high accuracy and reliability. In this work, we developed a strategy for the design of triple-mode immunoassays with the self-assemblies of three-in-one small molecules as signal reporters. Pyrroloquinoline quinone (PQQ), with a well-defined redox peak and excellent spectroscopic and fluorescent signals, was chosen as the signaling molecule. PQQ was coordinated with Cu²⁺ to form metal–organic nanoparticle as the signal label. Hexahistidine (His₆)-tagged recognition element (recombinant streptavidin) was attached to the Cu-PQQ surface through metal coordination interaction between the His₆ tag and the unsaturated metal site. The captured Cu-PQQ nanoparticle released a large number of PQQ molecules under an acidic condition, which could be simultaneously monitoring by electrochemical, UV-vis, and fluorescent techniques, thereby allowing for the development of triple-model immunoassays. The three methods were used to determine the concentration of carcinoembryonic antigen (CEA) with the detection limits of 0.01, 0.1, and 0.1 ng/mL, respectively. This strategy opens up a universal route for the preparation of multiple-model signal labels and the oriented immobilization of bioreceptors for molecular recognition. Full article

Background: Circadian rhythm disruption caused by shift work alters metabolic and hormonal pathways, which accelerates chronic disease onset, leading to decreased quality and quantity of life. This study aimed to determine whether a practical lifestyle intervention emphasizing nutrition timing and recovery habits could mitigate the metabolic and psychological effects of night-shift work. We conducted a randomized, open-label, crossover trial with two 8-week periods. Methods: Female healthcare workers (n = 13) aged 18–50 years with a body mass index (BMI) between 27 and 40 kg/m² and working predominantly night shifts (≥30 h/week for ≥6 months) were randomized. During the 8-week intervention phase, participants received daily text messages with guidance on food, sleep/rest, and physical activity and were provided with whey protein isolate powder and grain-based snack bars to consume during work shifts. The program targeted improved nutrient timing, adequate protein intake, and structured rest without formal exercise training, allowing evaluation of dietary and behavioral effects feasible for this population. Total caloric (~30 kcal/kg lean mass) and protein (2 g/kg lean mass) needs were measured, along with sleep/rest goals of 6–8 h/24 h. Primary outcome measures were change in visceral fat percentage (VF%) by DXA and mental/physical quality of life (RAND SF-12). Secondary outcomes included fasting triglycerides, ALT, blood glucose, LDL, actigraphy, and fecal microbiome. Mixed-design two-way ANOVA was conducted to assess the effects of group (immediate [IG] and delayed [DG]), time (baseline, 8-week crossover, and week 16), and Group $\times$ Time (GxT) interactions, and Bonferroni correction was applied to post hoc t-tests. Results: Eleven participants completed the study. Both groups increased dietary protein intake (p < 0.001), and a GxT interaction for VF% (p = 0.039) indicated DG reduced VF% to a greater extent (−0.335 ± 0.114% (p = 0.003) vs. 0.279 ± 0.543% (p = 0.158)). Mental and physical QOL, objectively measured physical activity and sleep, serum lipids and inflammatory markers, and fecal microbiota remained unchanged (p > 0.05 for all GxT). Conclusions: The findings suggest that targeted nutrition and recovery strategies can modestly improve dietary intake and visceral fat; however, consistent with prior work, interventions without structured exercise may be insufficient to reverse broader metabolic effects of circadian disruption. This trial was registered at ClinicalTrials.gov, identifier: NCT06158204, first registered: 28 November 2023. Full article

Background/Objectives: This study presents a secondary analysis from a previously published trial on strength training and postural correction in chronic headache patients. Here, we investigate changes in neck muscle function and tenderness, and their relationship with headache symptoms. Methods: A total of 22 headache patients from a single-arm open-label trial were included in this study to assess muscle function and tenderness. The maximum voluntary contraction of neck flexion and extension, shoulder elevation, and craniocervical flexion test were performed at baseline, week eight, and week 14. The extension/flexion ratio of the neck, the rate of force development, and the early rate of force development for shoulder elevation were calculated. Muscle tenderness was analyzed using the total tenderness score (TTS) and correlations between these outcomes and headache changes were explored. Results: After the intervention muscle tenderness significantly decreased (−5.6 ± 6.4; p < 0.001) and significant improvements in muscle function were observed. Correlations of muscle function showed a significant and moderate correlation between TTS and extension/flexion ratio (Spearman rho: 0.567, p = 0.014). Conclusions: The results indicate that strength training and postural correction improve muscle function and reduce pericranial tenderness in patients with chronic headaches. These findings suggest that muscle tenderness and extension/flexion ratio may be useful for monitoring exercise interventions focused on improving the strength and balance of the neck in patients with chronic headaches. Full article

Accurate brain tumor classification from MRI is often constrained by limited labeled data. We systematically compare conventional machine learning, deep learning, and few-shot learning (FSL) for four classes (glioma, meningioma, pituitary, no tumor) using a standardized pipeline. Models are trained on the Kaggle Brain Tumor MRI Dataset and evaluated across dataset regimes (100%→10%). We further test generalization on BraTS and quantify robustness to resolution changes, acquisition noise, and modality shift (T1→FLAIR). To support clinical trust, we add visual explanations (Grad-CAM/saliency) and report per-class results (confusion matrices). A fairness-aligned protocol (shared splits, optimizer, early stopping) and a complexity analysis (parameters/FLOPs) enable balanced comparison. With full data, Convolutional Neural Networks (CNNs)/Residual Networks (ResNets) perform strongly but degrade with 10% data; Model-Agnostic Meta-Learning (MAML) retains competitive performance (AUC-ROC ≥ 0.9595 at 10%). Under cross-dataset validation (BraTS), FSL—particularly MAML—shows smaller performance drops than CNN/ResNet. Variability tests reveal FSL’s relative robustness to down-resolution and noise, although modality shift remains challenging for all models. Interpretability maps confirm correct activations on tumor regions in true positives and explain systematic errors (e.g., “no tumor”→pituitary). Conclusion: FSL provides accurate, data-efficient, and comparatively robust tumor classification under distribution shift. The added per-class analysis, interpretability, and complexity metrics strengthen clinical relevance and transparency. Full article

Reinforcement Learning from Human Feedback (RLHF) produces powerful instruction-following models but relies on a preference-labeling process that is both costly and slow. An effective alternative, Reinforcement Learning from AI Feedback (RLAIF), uses large language models as teachers for relabeling; however, this introduces substantial label noise. In our setting, we found that AI teachers flipped approximately 50% of the original human preferences on the dataset, a condition that degrades the performance of standard direct preference optimization (DPO). We propose noise-robust DPO (nrDPO) and nrDPO-gated, two drop-in variants that make DPO resilient to noisy preferences. nrDPO reweights each pair by (i) a margin-confidence term from a frozen reference policy (base or SFT), (ii) a context-stability term that penalizes preferences that change under truncated histories, and (iii) a length correction to curb verbosity bias. nrDPO-gated further filters low-confidence pairs via a simple threshold on the reference margin. On a dataset with heavy synthetic noise (30% flips), nrDPO-gated improves the preference accuracy by +3.8% over vanilla DPO; in a realistic RLAIF setting, nrDPO-gated is the only configuration that recovers competitive alignment, reaching ≈60% on a 5k relabeled set (vs. ≈49–50% for vanilla DPO) and approaching RLHF baselines. Full article

Gait recognition, as a biometric technology with great potential, has been widely applied in numerous fields due to its unique advantages. Through in-depth research and the creation of in-the-wild gait datasets, gait recognition technology is progressively extending from laboratory settings to complex real-world scenarios, achieving notable advancements. However, the complexity of annotating gait data inevitably leads to labeling errors, known as noisy labels, which are one of the reasons for the suboptimal performance of in-the-wild gait recognition. To address these issues, this paper explores noisy label learning for in-the-wild gait recognition for the first time. We propose a plug-and-play gait recognition framework named Dynamic Noise Label Correction Network (DNLC). Specifically, it consists of two main parts: the dynamic class-center feature library and the label correction module, which can automatically identify and correct noisy labels based on the class-center feature library. In addition, we introduce the two-stage augmentation strategy to increase the diversity of the data and help reduce the impact of noisy labels. We integrated our proposed framework into five existing gait models and conducted extensive experiments on two widely used gait recognition datasets: Gait3D and CCPG. The results show that our framework increased the average Rank-1 accuracy of five methods by 10.03% and 6.45% on the Gait3D and CCPG datasets, respectively. These findings demonstrate the superior performance of our method. Full article

Accurate long-term tracking of individual cattle is essential for precision livestock farming but remains challenging due to occlusions, posture variability, and identity drift in free-range environments. We propose a multi-camera tracking framework that combines bird’s-eye-view (BEV) trajectory matching with cattle face recognition to ensure identity preservation across long video sequences. A large-scale dataset was collected from five synchronized 4K cameras in a commercial barn, capturing both full-body movements and frontal facial views. The system employs center point detection and BEV projection for cross-view trajectory association, while periodic face recognition during feeding refreshes identity assignments and corrects errors. Evaluations on a two-day dataset of more than 600,000 images demonstrate robust performance, with an AssPr of 84.481% and a LocA score of 78.836%. The framework outperforms baseline trajectory matching methods, maintaining identity consistency under dense crowding and noisy labels. These results demonstrate a practical and scalable solution for automated cattle monitoring, advancing data-driven livestock management and welfare. Full article

Background: The inability to assign a protein content claim (PCC) to plant foods may impede efforts from Canada’s Food Guide to increase consumption of plant protein. A systematic application of PCC frameworks from other regions to Canadian nutrition surveillance data would be useful to model potential effects of PCC regulations on the nutrient intake, protein quality, and corrected protein intake of diets. Methods: Plant food groups that qualified for a PCC within the Canadian Nutrient File according to regulations from Canada, the United States (US), Australia and New Zealand (ANZ), and the European Union (EU) were identified. Adults (≥19 years) (n = 11,817) from The Canadian Community Health Survey (2015) who consumed ≥1 plant food qualifying for a PCC in each region were allocated to the corresponding PCC group. The effects of Canadian PCC regulations on the protein quantity, quality (Digestible Indispensable Amino Acid Score, DIAAS), and nutrient intakes of Canadian diets in adults were compared to PCC groups from other regions. Results: Substantially more individuals were consumers of plant-based protein foods, using the ANZ and the EU PCC regulations, compared to the Canadian and US PCC groups. There were no differences in uncorrected protein intake across PCC groups. All DIAAS values were >0.94, and corrected protein intakes were >74–89 g/day or 16%E across PCC groups. Non-consumers of plant foods eligible for a PCC had corrected protein intakes that ranged between 68 and 78 g/d or 17%E. Generally, consumers of plant foods eligible for a PCC in the US, ANZ, and EU, or both Canada and the US/ANZ/EU, had higher intakes of positive nutrients, such as fibre, calcium, iron, magnesium, and zinc (p < 0.05) and lower saturated fat. Conclusions: Less restrictive regulatory frameworks for PCC used in ANZ and the EU did not substantially affect protein intake or the protein quality of Canadian diets in adults. These results suggest that more inclusive regulatory frameworks for protein PCCs could support increased intake of food sources of plant proteins in alignment with Canada’s Food Guide. Full article

Intravital two-photon microscopy enables deep-tissue imaging of subcellular structures in live animals, but its original spatial resolution and image quality are limited by scattering, motion, and low signal-to-noise ratios. To address these challenges, we used a combination of tissue stabilization, denoising methods, and motion correction, together with resolution enhancement algorithms, including enhanced Super-Resolution Radial Fluctuations (eSRRF) and deconvolution, to acquire high-fidelity time-lapse images of internal organs. We applied this imaging pipeline to image genetically labeled mitochondria in vivo, in Dendra2 mice. Our results demonstrate that the eSRRF-combined method, compared to other evaluated algorithms, significantly shows improved spatial resolution and mitochondrial structure visualization, while each method exhibiting distinct strengths in terms of noise tolerance, edge preservation, and computational efficiency. These findings provide a practical framework for selecting enhancement strategies in intravital imaging studies targeting dynamic subcellular processes. Full article

Background/Objectives: Neuroendocrine neoplasms (NENs) are uncommon neoplasms. Grading informs the prognosis and treatment decision of NENs and is determined by cell proliferation, which is measured by mitotic count and Ki-67 index. These measurements present challenges for pathologists as they suffer inter- and intra-observer variability and are cumbersome to quantify. To address these challenges, we developed a machine learning pipeline for identifying tumor areas, proliferating cells, and grading NENs. Methods: Our study includes 385 samples of gastroenteropancreatic NENs from across British Columbia with two stains (247 H&E and 138 Ki-67 images). Labels for these cases are at the patient-level, and there are 186 patients. We systematically investigated three settings for our study: H&E, H&E with Ki-67, and pathologist-reviewed and corrected cases. Results: Our H&E framework achieved a three-fold balanced accuracy of 77.5% in NEN grading. The H&E with Ki-67 framework yields a performance improvement to 83.0% on grading. We provide survival and multivariate analysis with a c-index of 0.65. Grade 1 NENs misclassified by the model were reviewed by a pathologist to assess reasons. Analysis of our AI-graded NENs for the subset of pathologist-assessed G1s demonstrated a significant (p-value = 0.007) survival difference amongst samples the algorithm assigned to a higher grade (n = 20; median survival 4.22 years) compared to concordant G1 samples (n = 60; median survival 10.13 years). Conclusions: Our model identifies NEN grades with high accuracy and identified some grade 1 tumors as prognostically unique, suggesting potential improvements to standard grading. Further studies are needed to determine if this discordant group is a different clinical entity. Full article

Weakly supervised semantic segmentation (WSSS) aims to learn pixel-level semantic concepts from image-level class labels. Due to the simplicity and efficiency of the training pipeline, end-to-end WSSS has received significant attention from the research community. However, the coarse nature of pseudo-label regions remains one of the primary bottlenecks limiting the performance of such methods. To address this issue, we propose class-guided enhanced pseudo-labeling (CEP), a method designed to generate high-quality pseudo-labels for end-to-end WSSS frameworks. Our approach leverages pretrained foundation models, such as contrastive language-image pre-training (CLIP) and segment anything model (SAM), to enhance pseudo-label quality. Specifically, following the pseudo-label generation pipeline, we introduce two key components: a Skip-CAM module and a pseudo-label refinement module. The Skip-CAM module enriches feature representations by introducing skip connections from multiple blocks of CLIP, thereby improving the quality of localization maps. The refinement module then utilizes SAM to refine and correct the pseudo-labels based on the initial class-specific regions derived from the localization maps. Experimental results demonstrate that our method surpasses the state-of-the-art end-to-end approaches as well as many multi-stage competitors. Full article

The sit-and-reach test is a common stretching exercise suitable for adolescents, aimed at improving joint flexibility and somatic neural control, and has become a mandatory item in China’s student physical fitness assessments. However, many students tend to perform incorrect postures during their practice, which may lead to sports injuries such as muscle strains if sustained over time. To address this issue, this paper proposes a Ghost-ST-GCN model for judging the correctness of the sit-and-reach pose. The model first requires detecting seven body keypoints. Leveraging a publicly available labeled keypoint dataset and unlabeled sit-and-reach videos, these keypoints are acquired through the proposed self-train method using the BlazePose network. Subsequently, the keypoints are fed into the Ghost-ST-GCN model, which consists of nine stacked GCN-TCN blocks. Critically, each GCN-TCN layer is embedded with a ghost layer to enhance efficiency. Finally, a classification layer determines the movement’s correctness. Experimental results demonstrate that the self-train method significantly improves the annotation accuracy of the seven keypoints; the integration of ghost layers streamlines the overall detection model; and the system achieves an action detection accuracy of 85.20% for the sit-and-reach exercise, with a response latency of less than 1 s. This approach is highly suitable for guiding adolescents to standardize their movements during independent sit-and-reach practice. Full article

Due to the local similarities between different facial expressions and the subjective influences of annotators, large-scale facial expression datasets contain significant label noise. Recognition-based noisy labels are a key challenge in the field of deep facial expression recognition (FER). Based on this, this paper proposes a simple and effective attention-based fusion consistency network (AFCN), which suppresses the impact of uncertainty and prevents deep networks from overemphasising local features. Specifically, the AFCN comprises four modules: a sample certainty analysis module, a label correction module, an attention fusion module, and a fusion consistency learning module. Among these, the sample certainty analysis module is designed to calculate the certainty of each input facial expression image; the label correction module re-labels samples with low certainty based on the model’s prediction results; the attention fusion module identifies all possible key regions of facial expressions and fuses them; the fusion consistency learning module constrains the model to maintain consistency between the regions of interest for the actual labels of facial expressions and the fusion of all possible key regions of facial expressions. This guides the model to perceive and learn global facial expression features and prevents it from incorrectly classifying expressions based solely on local features associated with noisy labels. Experiments are conducted on multiple noisy datasets to validate the effectiveness of the proposed method. The experimental results illustrate that the proposed method outperforms current state-of-the-art methods, achieving a 3.03% accuracy improvement on the 30% noisy RAF-DB dataset in particular. Full article

In semantic segmentation tasks, large kernels and Atrous convolution have been utilized to increase the receptive field, enabling models to achieve competitive performance with fewer parameters. However, due to the fixed size of kernel functions, networks incorporating large convolutional kernels are limited in adaptively capturing multi-scale features and fail to effectively leverage global contextual information. To address this issue, we combine Atrous convolution with large kernel convolution, using different dilation rates to compensate for the single-scale receptive field limitation of large kernels. Simultaneously, we employ a dynamic selection mechanism to adaptively highlight the most important spatial features based on global information. Additionally, to enhance the model’s ability to fit the true label distribution, we propose a Multi-Scale Contextual Noise Transfer Matrix (NTM), which uses high-order consistency information from neighborhood representations to estimate NTM and correct supervision signals, thereby improving the model’s generalization capability. Extensive experiments conducted on Cityscapes, ADE20K, and COCO-Stuff-10K demonstrate that this approach achieves a new state-of-the-art balance between speed and accuracy. Specifically, LKNTNet achieves 80.05% mIoU on Cityscapes with an inference speed of 80.7 FPS and 42.7% mIoU on ADE20K with an inference speed of 143.6 FPS. Full article