APAU-Net: Adaptive Prior-Aware U-Net Text-Line Segmentation for Historical Documents

Text-line segmentation in historical manuscripts remains challenging due to degradation, overlapping strokes, and extreme data scarcity. We propose APAU-Net, a two-stage cascade architecture that explicitly optimizes Line Intersection-over-Union (Line IU) via learned anisotropic Gaussian priors. Stage 1 predicts topology-aware ellipsoidal priors from low resolution grayscale images using connected-component analysis and moment-based ellipse fitting. Stage 2 refines these priors at full resolution through a residual U-Net with adaptive per-pixel weighting and gated fusion. We evaluated APAU-Net on the challenging U-DIADS-TL (84 images, only 3 training pages per manuscript) and DIVA-HisDB benchmarks. It achieves an average Line Intersection-over-Union (Line IU) of 94.3% (+9.8 pp over plain U-Net) and outperforms recent few-shot baselines by up to +24 pp on the most degraded Syriac subset. Ablation confirms the anisotropic prior contributes ~9-12 pp to Line IU under severe data scarcity. The source code for the proposed method is openly available on GitHub at \href{https://github.com/aminebeg/APAUNet}{https://github.com/aminebeg/APAUNet}