Cluster-First Labelling: An Automated Pipeline for Segmentation and Morphological Clustering in Histology Whole Slide Images

Muhammad Haseeb Ahmad; Sharmila Rajendran; Damion Young; Jon Mason

q-bio.QMApr 10, 2026

Cluster-First Labelling: An Automated Pipeline for Segmentation and Morphological Clustering in Histology Whole Slide Images

Authors: Muhammad Haseeb Ahmad, Sharmila Rajendran, Damion Young, Jon Mason

Organizations: Department of Engineering Science, University of Oxford, United Kingdom · Department of Physiology, Anatomy and Genetics, University of Oxford, United Kingdom · Medical Sciences Division, University of Oxford, United Kingdom

Abstract

Labelling tissue components in histology whole slide images (WSIs) is prohibitively labour-intensive: a single slide may contain tens of thousands of structures--cells, nuclei, and other morphologically distinct objects--each requiring manual boundary delineation and classification. We present a cloudnative, end-to-end pipeline that automates this process through a cluster-first paradigm. Our system tiles WSIs, filters out tiles deemed unlikely to contain valuable information, segments tissue components with Cellpose-SAM (including cells, nuclei, and other morphologically similar structures), extracts neural embeddings via a pretrained ResNet-50, reduces dimensionality with UMAP, and groups morphologically similar objects using DBSCAN clustering. Under this paradigm, a human annotator labels representative clusters rather than individual objects, reducing annotation effort by orders of magnitude. We evaluate the pipeline on 3,696 tissue components across 13 diverse tissue types from three species (human, rat, rabbit), measuring how well unsupervised clusters align with independent human labels via per-tile Hungarian-algorithm matching. Our system achieves a weighted cluster-label alignment accuracy of 96.8%, with 7 of 13 tissue types reaching perfect agreement. The pipeline, a companion labelling web application, and all evaluation code are released as open-source software.

Figures & tables

Figure 1: Pipeline architecture. WSI slides ( .ndpi ) are tiled into $512\times 512$ patches, quality-filtered, segmented with Cellpose-SAM, embedded with ResNet-50, reduced via UMAP, and clustered with DBSCAN. Optional downstream stages annotate images, extract representative tiles per cluster, and invoke a multimodal LLM for experimental classification.

Figure 2: Labelling application interface. The central panel shows a $512\times 512$ histology tile with interactive cell polygons. The sidebar provides tile navigation, label assignment controls, per-tile and batch progress, and visual customisation settings.

Batch	Cells	Tiles	Human	Model	Acc. (%)
			clusters	clusters
Batch 1	1,229	5	2	1	94.9
Batch 2	1,117	15	3	2	97.7
Batch 3	1,350	9	2	1	97.8
Total	3,696	29	—	—	96.8

Table 1: Batch-level evaluation results. Accuracy is the weighted Hungarian-aligned accuracy across all tiles in each batch.

Tissue type	Species	Cells	Tiles	Acc. (%)
Appendix	Human	662	1	91.4
Ovary	Human	587	2	99.7
Compact bone	Human	106	11	84.0
Skeletal muscle	Human	219	3	84.0
Pancreas	Human	28	1	100.0
Kidney	Rat	328	2	97.9

Table 2: Per-tissue-type evaluation. Accuracy is the weighted Hungarian-aligned accuracy over all tiles from that slide, pooled across batches.

Figure 3: Evaluation interface showing batch-level accuracy (97.67%) and per-tile confusion matrices. Rows represent human labels; columns represent Hungarian-aligned model clusters. Green cells indicate correct assignments.

Figure 4: Post-alignment tile comparison. Cell polygons are colour-coded: green = human and aligned-model labels agree; red = disagree. Labels are shown as human / aligned-model pairs. Bottom panel: rabbit femur tile with 207 cells, 206 correct (99.5% accuracy).

Whole-Slide Image Analysis

Whole-slide image analysis uses AI to interpret gigapixel pathology slides for diagnosis, prognosis, tissue mapping, and report generation.

See all 214 papers

Slide-Level Active Learning Reduces Annotation Burden in H&E images
Jul 10, 2026Mahsa Vali, Zhilong Weng, Noémie Moreaua +2Computational Pathology Histopathology Image Segmentation
Institute for Biomedical Informatics, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany · Center for Molecular Medicine Cologne (CMMC), Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany · Faculty of Mathematics and Natural Sciences, University of Cologne, Cologne, Germany +3
SlideMix: Enhancing Whole Slide Image Analysis via Multimodal Shuffling
Aug 31, 2026Chad Wong, Sicheng Chen, Tianyi Zhang +4Data Augmentation Histopathology Image Classification
1Nhu Department of Electrical Engineering and Computer Science, University of California, Irvine · Department of Electrical & Computer Engineering, National University of Singapore · 3PuzzleLogic Pte Ltd, Singapore 229594, Singapore
Nonparametric Distribution Matching for Self-Supervised Whole-Slide Image Condensation
Sep 30, 2026Duong M. Nguyen, Trong Nghia Hoang, Hang Thi Nguyen +3Computational Pathology Distribution Matching
University of Illinois Urbana-Champaign · Washington State University · Hanoi Medical University +2

Jul 10, 2026J/K move · Enter open · S save

Slide-Level Active Learning Reduces Annotation Burden in H&E images

Mahsa Vali, Zhilong Weng, Noémie Moreaua, Yuri Tolkach+1

Institute for Biomedical Informatics, Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany · Center for Molecular Medicine Cologne (CMMC), Faculty of Medicine and University Hospital Cologne, University of Cologne, Cologne, Germany · Faculty of Mathematics and Natural Sciences, University of Cologne, Cologne, Germany+3

Computational Pathology Histopathology Image Segmentation Medical Image Segmentation Whole-Slide Image Analysis

Open paper AskPDF arXiv

Deep learning-based segmentation of histopathology whole-slide images (WSIs) requires large amounts of pixel-level annotations, which are costly and time-consuming to obtain. Active learning (AL) has been proposed to reduce this effort, but existing methods exhibit three key limitations. Uncertainty estimation is unreliable on partially annotated WSIs, patch-level acquisition is inconsistent with slide-level annotation workflows, and class imbalance in multi-class settings is not explicitly addressed. To address these challenges, we propose SHAL (Slide-level Hybrid Active Learning), a patient-level AL framework for annotation-efficient multi-class histopathology segmentation. SHAL integrates three complementary components: a foreground-aware strategy that suppresses bias from unlabeled background regions, a stage-adaptive mechanism that hybridizes predictive entropy and epistemic uncertainty across learning stages, and a class-aware strategy that prioritizes diagnostically relevant tissue classes. SHAL is evaluated on the TCGA colorectal cancer dataset. It achieves the highest Macro Dice at the full annotation budget (0.846) and reaches Dice greater than or equal to 0.80 using only 26 percent of the budget (50 of 190 slides), whereas competing methods reach this threshold only at 37 percent (70 slides). Across five independent external cohorts, SHAL attains the highest mean external Macro Dice (0.815) and the smallest internal-to-external generalization gap among all methods (0.025 at Round 3 and 0.026 at the full budget). The results indicate that patient-level hybrid uncertainty acquisition reduces annotation cost without sacrificing cross-domain generalization in computational pathology.

Abstract

Whole-Slide Image Analysis

Explore similar work

Slide-Level Active Learning Reduces Annotation Burden in H&E images

SlideMix: Enhancing Whole Slide Image Analysis via Multimodal Shuffling

Nonparametric Distribution Matching for Self-Supervised Whole-Slide Image Condensation