{"id":"arxiv:2610.31902","kind":"preprint","title":"Automated Nucleus Segmentation in Histology Images: An Attention U-Net Approach with Comprehensive Preprocessing and Inference Pipelines","authors":["Akshat G","Divyansh Gupta","Tusar Kanti Mishra"],"author_count":3,"year":2026,"published":"2026-10-22","venue":null,"cited_by":null,"doi":null,"arxiv_id":"2610.31902","nct_id":null,"source_url":"https://arxiv.org/abs/2610.31902","archive_url":"https://space-kitty.com/papers/arxiv:2610.31902","updated":"2026-10-22","pmid":null,"pmcid":null,"openalex_id":null,"pdf_url":"https://arxiv.org/pdf/2610.31902","abstract":"Objectives: In the area of digital pathology, nucleus segmentation is an essential step for quantitative analysis; however, high nuclear density and staining heterogeneity in histology images pose a great challenge. Methods: This study implements a lightweight Attention U-Net model on the MoNuSeg 2018 dataset to overcome these challenges. The proposed pipeline integrates comprehensive image and mask preprocessing, including stain deconvolution and multi-pass mask smoothing, with a high-fidelity overlapping patch-stitching strategy during inference on high-resolution images. Results: The model achieved an Intersection over Union (IoU) score of 0.8745 on full-resolution, unseen test images. This result demonstrates the effectiveness of attention mechanisms in enhancing feature localization and validates the stitching methodology for generating seamless and accurate segmentation maps. Conclusion: Attention U-Net is an effective and competitive method for automated, scalable nucleus delineation in challenging clinical contexts.","abstract_license":"CC0-1.0 (arXiv descriptive metadata)","license":null,"is_open_access":false,"retracted":false,"peer_reviewed":false,"categories":["cs.CV"],"details":null,"provenance":"arxiv","attribution":"Abstract: arXiv descriptive metadata (CC0). Thank you to arXiv for use of its open access interoperability. This service was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.","topics":[],"cited_in_dossiers":[]}