• 学术搜索
  • 科研智能体
    • Research Labs
    • AI 阅读
    • AI 文库
    • 深度研究
    • 学者亮点
  • 学术资源
    • AI2000
    • 期刊/会议
    • 学者库
    • 学术API
      MCP&Skill
    • 溯源树
    • 数据集
  • 知识沉淀
    • 学术空间
订阅小程序
旧版功能
aminer vip
开通会员低至0.73元/天
一次搞定AI科研
立即登录
  • English
  • 联系方式
    Artificial Intelligence in Health

    Artificial Intelligence in Health

    JournalISSN 3041-0894eISSN 3029-2387

    年发文量

    研究主题

    论文(185)

    排序
    1A Study on Temporal Consistency and Comparative Performance of a NASNet-based Deep Learning Model for Endoscopic Video Denoising
    Amulya H. Gopalakrishna, Gopalakrishna Kanathur

    Image processing plays an important role in endoscopic diagnosis; noise introduced by low illumination, sensor limitations, motion artifacts, and endoscope movement within the human body can significantly reduce visual clarity. Image and video quality affect the visibility of fine structures, making diagnosis difficult. Deep learning-based methods have shown promising results, but many existing models struggle to maintain temporal consistency across consecutive frames. This paper presents an enhanced video denoising framework based on a neural architecture search network (NASNet) model with a channel attention mechanism. The proposed method was optimized using a genetically modified jellyfish search optimization (GM-JFSO) algorithm. The proposed framework was evaluated using the publicly available Hyper–Kvasir dataset. It aims to preserve spatial image quality and temporal consistency across consecutive video frames by effectively reducing noise while retaining structural details. Comparative evaluation against NASNet, GoogLeNet, DenseNet, and ResNet showed that the proposed method outperformed them, achieving a peak signal-to-noise ratio of 49.07 dB, a structural similarity index measure (SSIM) of 0.938, a mean squared error of 0.584, and a temporal SSIM of 0.8313 across the evaluated sequence. The GM-JFSO algorithm improves the proposed model’s performance through hyperparameter optimization, while the channel attention mechanism enhances feature representation and noise suppression. The experiment shows the effectiveness of integrating the optimization technique with the deep learning model for video denoising and highlights the importance of evaluating temporal consistency in medical imaging applications.

    2026
    引用
    AI阅读
    加入学术空间
    2Verification Feedback and Self-Revision for Omissions in Clinical Recommendations by Large Language Models: A Paired Evaluation
    Sanjay Basu, Sadiq Patel, Parth Sheth, John Morgan, Rajaie Batniji

    Large language models (LLMs) in clinical decision support can omit required actions (omission) or recommend wrong ones (commission). Verification layers check responses, and models can revise without feedback. Based on 176 test cases, we delivered the first responses of three LLMs (GPT-6 Astra, Gemini 3.1 Pro, Claude Opus 5.5) in six versions: no feedback, the full findings of a rule-and-agent verification framework, or alerts calibrated to a 10% false-alert budget, each appended or returned to the model for revision. Three LLM judges scored the share of high-grade (high-certainty) items on LLM-generated checklists that each text omitted; physicians reviewed samples. Revision without feedback (self-revision) lowered omission from 8.1% to 7.1% (−0.9 percentage points; 95% confidence interval [CI]: −1.7 to −0.1). In post hoc comparisons, revision with the full findings omitted 2.2 percentage points more than self-revision (95% CI: 1.3 to 3.2), or 0.8 (95% CI: −0.1 to 1.8) without the Gemini 3.1 Pro revisions that replaced content with status phrases. Revision with the alerts, which reached 56 of 528 first responses, did not differ from self-revision (0.5 percentage points; 95% CI: −0.3 to 1.3). Appended, the full findings changed judged omission by −0.6 percentage points (95% CI: −1.20 to −0.04). Based on 354 responses to physician-reviewed commission probes, both deliveries of the full findings lowered commission, mostly citations to nonexistent guidelines, from 6.8% to 3.4% (appended) and 4.2% (revised); with hedged commission counted, only revision lowered it (post hoc). Revision with verification feedback did not reduce omission beyond self-revision, which evaluations should include as a baseline.

    2026
    引用
    AI阅读
    加入学术空间
    3The Calibration and Energy Cost of Tabular Foundation Models in Clinical Prediction
    Muhammad Ahmad Zia, Muhammad Muzaffar, Yosef Hasan Jbara, Bassam Elzaghmouri, Marwan Abu-Zanona, Danish Mahmood Khan

    Artificial intelligence (AI) is increasingly used in healthcare decision-making, where predictive models must accurately distinguish patient risk, provide reliable probabilities, and remain computationally feasible for routine clinical deployment. Despite the emergence of tabular foundation models, evaluation has focused predominantly on predictive accuracy, providing limited evidence of their practical advantages in healthcare. This study evaluates the tabular foundation model TabPFN and calibrated gradient-boosted decision tree (GBDT) ensembles across multiple publicly available clinical prediction datasets. The calibration comparison followed a pre-registration fixing the equivalence margin, an inclusion gate that removed models that had not learned, and an explicit falsifier. Energy was read from processor power counters, each figure tagged by hardware, configuration, and batch regime. TabPFN was better calibrated on all five datasets across four calibration metrics. A saturated dataset carries no discriminating weight, so the confirmatory conclusion rests on the other four. Rejecting the pre-registered equivalence hypothesis indicates a calibration difference favoring TabPFN, rather than equivalence, although the gap was small (mean adaptive expected calibration error = 0.011). TabPFN showed better discrimination on four of five datasets, with a mean AUROC difference of 0.013 and intervals that exclude zero. Decision-curve analysis showed a small positive net benefit at most thresholds. Its modest gains in calibration, discrimination, and net benefit came at substantially higher computational cost. Tree models were more energy-efficient beyond approximately six predictions; per-prediction cost jumped 121-fold with a single configuration flag, and a validated cost model showed inference growing steeply with context size, making CPU-only deployment impractical at clinical scale. Thus, TabPFN’s advantage is narrow and most justified when calibrated probabilities carry clinical weight, volume is low, and an accelerator is available, where cost is characterized by timing rather than measured power.

    2026
    引用
    AI阅读
    加入学术空间
    4Artificial Intelligence in Skin Cancer Diagnosis and Prognosis: A Comprehensive Narrative Review of Current Applications and Future Perspectives
    Harpreet Singh, Rajdeep Singh, Pinki Kushwaha, Jatin Agarwal, Sagar Varshney,Arvind Kumar,Arun Kumar Mishra,Hitesh Chopra,Shivani Chopra, Tabarak Malik

    Skin cancer is among the most prevalent malignancies worldwide, and early, accurate diagnosis is crucial for improving patient outcomes. Recent advances in artificial intelligence (AI), particularly machine learning and deep learning, have shown substantial potential to enhance skin cancer detection, classification, and prognostic assessment. This review provides a comprehensive synthesis of current AI-based approaches for melanoma and non-melanoma skin cancers, highlighting methodological innovations and clinical applications. Literature from 2015 to 2025 was screened from PubMed/MEDLINE, Scopus, IEEE Xplore, and ScienceDirect, focusing on peer-reviewed studies reporting AI-driven diagnostic, classification, or prognostic outcomes. Deep learning models, especially convolutional neural networks, demonstrated high diagnostic performance in image-based skin cancer detection, often comparable to experienced dermatologists. AI has also shown promise in lesion segmentation, risk stratification, and prognostic modeling. However, challenges remain, including class imbalance, underrepresentation of darker skin tones, limited external validation, algorithmic opacity, and integration into clinical workflows. To enable broader clinical adoption, future research should prioritize diverse multicenter datasets, explainable AI systems, multimodal data integration, and prospective clinical validation studies. Overall, AI technologies offer significant potential to improve the accuracy and efficiency of skin cancer diagnosis and prognosis, but their translation into routine dermatological practice requires careful attention to reliability, equity, and interpretability.

    2026
    引用
    AI阅读
    加入学术空间
    5A Hybrid Feature Selection and Ensemble Learning Model for Heart Disease Diagnosis
    Anurag Tripathi, Rahul Vyas, Gaurav Dwivedi, Lalit Kumar Tripathi, Ajai Kumar Maurya, Ali Zaheer Agha

    Heart disease remains one of the leading causes of mortality worldwide, making early and accurate diagnosis essential for improving patient outcomes. This study proposes a hybrid feature selection and ensemble learning framework for heart disease diagnosis using a publicly available dataset containing 303 records and 13 clinical attributes. The preprocessing pipeline included missing-value imputation, one-hot encoding of categorical variables, and minimum–maximum normalization. Support vector machine–recursive feature elimination (SVM–RFE) was used to identify the most informative clinical features and remove redundant attributes. Multiple machine learning classifiers, including logistic regression, k-nearest neighbors, SVMs, decision trees, random forests, and extreme gradient boosting (XGBoost), were evaluated. Bayesian optimization was applied to tune model hyperparameters, and performance was assessed using an 80:20 train–test split together with stratified five-fold cross-validation. The proposed ensemble model achieved 93.62% accuracy, 94.08% precision, 93.11% recall, 93.58% F1-score, and 0.96 receiver operating characteristic area under the curve. The selected features included chest pain type, ST-slope, exercise-induced angina, maximum heart rate, and age, all of which are clinically meaningful predictors of cardiovascular risk. The results show that combining SVM-RFE, hyperparameter optimization, and ensemble learning improved predictive performance, generalization, and interpretability. The proposed framework can support early heart disease screening and clinical decision-making.

    2026
    引用
    AI阅读
    加入学术空间
    立即登录,查看全部 185 篇论文

    高被引作者

    作者引用发文
    Xiaojing Xu592
    Büşra Tuğçe Susam592
    Kenneth D. Craig592
    Virginia R. De Sa592
    Murat Akçakaya592
    Matthew S. Goodwin592
    Jeannie S Huang592
    Damaris Diaz592
    Alvin Shi281
    Alex Chan281

    高产作者

    作者引用发文
    Xiaojing Xu592
    Virginia R. De Sa592
    Fatma Indriani02
    Saurav Mallik02
    Jeannie S Huang592
    Carly Hudson02
    Yianna Vovides22
    Hswen Yulin12
    Radityo Adi Nugroho02
    Aki Härmä12

    相关期刊

    暂无数据
    暂无数据