July 2026 in “International Journal of Advanced Research in Science Communication and Technology” In this study, the BaldGraphFormer framework, integrating visual and clinical data, outperformed unimodal baselines in early-stage androgenetic alopecia detection, achieving an F1-score of 97.62% and macro-average AUC of 0.992, suggesting its potential to support dermatological decision-making and early intervention.
4 citations
,
April 2024 in “Complex & Intelligent Systems” This study introduced a single-stage network using large kernel attention that effectively restores high-resolution images by capturing both global and local details, reducing parameters and improving processing speed.
5 citations
,
January 2009 in “Elsevier eBooks” This chapter discusses Bayesian networks, their properties, and how they enable probabilistic inference with continuous variables, but it reports no new findings.
7 citations
,
April 2023 in “Frontiers in immunology” This source reports that a network meta-analysis found oral baricitinib and ruxolitinib to be effective and safe for treating alopecia areata, significantly improving response rates over placebos and showing fewer adverse effects than steroid treatments, though precise dosing needs further study.
April 2024 in “Frontiers in pharmacology” This study compared JAK inhibitors for moderate-to-severe alopecia areata and found brepocitinib 30mg and deuruxolitinib 12mg to be most effective, though deuruxolitinib may cause more adverse events. Further research is needed for long-term efficacy and safety.