Existing approaches to building line-level Arabic handwritten-text-recognition (HTR) training data either rely on fully manual annotation, which does not scale, or on automatic OCR-to-reference alignment methods not yet extended to multi-script, two-zone (main-plus-margin) manuscript layouts with a provable correctness guarantee. We present RefLAM (Reference-grounded Line Annotation for Manuscripts), a pipeline converting manuscript page images and clean transcriptions into validated, line-level ground truth without sacrificing human oversight. RefLAM couples a deep-learning page-segmentation model with a multimodal large language model (MLLM) for structured OCR and a diacritic-agnostic fuzzy alignment engine that grounds each OCR line in a contiguous span of the reference text, with a character-level confidence score in $[0,100]$. A perfect score is provably equivalent to character-for-character identity of the normalised strings (the Confidence-100 rule), verified with no counterexample across the released corpus. A reviewer can thus trust a perfect score, confirming most lines at a glance rather than retyping them, so annotation becomes triaged, with attention concentrated on uncertain alignments. Across 7 fully page-validated books we measured a 75$\times$ throughput gain over manual annotation (3,000 vs. 40 lines/hr); applying the same guarantee to 7 further books, we retained 16,533 confidence-100 main-text lines within one week, excluding sub-100 lines rather than manually correcting them. Using RefLAM, we release AraMS-28k: 14 historical Arabic manuscript books, 3,043 pages, and 27,971 main-text and 629 margin-line annotations with bounding boxes, layout labels, and insertion anchors for 191 margin entries (30.4%). We also finetune Muharaf-pretrained baselines (including HATFormer) on AraMS-28k and report CER results confirming its practical utility for downstream HTR training.
Federated learning (FL) enables multi-institutional collaboration in medical imaging while preserving patient privacy, yet its fairness landscape remains fragmented: existing methods predominantly address either
collaboration fairness
(equitable performance across institutions) or
group fairness
(equitable outcomes across demographic subgroups), but rarely both. In this systematic review, we adopt
dual fairness
—the joint satisfaction of both dimensions—as the analytical lens for organizing and critically evaluating this landscape. Following the PRISMA 2020 guidelines, we analyze 132 publications and classify fairness-aware FL methods through a three-dimensional taxonomy: client-side, server-side, and communication-based approaches. Among the 20 fairness-aware or fairness-adapted FL methods catalogued, only three partially address both dimensions, and none provides provable joint guarantees under clinically realistic conditions. Our critical analysis identifies three fundamental challenges: the Local–Global Pareto Frontier Conflict, in which collaboration and group fairness gradients in the accuracy space can exceed
$$150^{\circ }$$
under sufficiently asymmetric demographic imbalance (e.g.,
$$\pi _1 \ge 0.8$$
); the Privacy–Fairness Compounding Effect, through which differential privacy mechanisms disproportionately suppress minority gradient signals; and the risk of pseudo-fairness, whereby equipment–demographic confounding masks genuine algorithmic discrimination. We further outline a seven-direction research roadmap. To the best of our knowledge, this constitutes the first systematic review to formally analyze the gradient-level conflict between collaboration fairness and group fairness in federated medical imaging, while also providing a structured causal analysis of equipment–demographic confounding, offering both a critical synthesis and actionable directions toward equitable AI-assisted healthcare.
