Multi-Stage Fine-Tuning of Pathology Foundation Models with Head-Diverse Ensembling for White Blood Cell Classification

Antony Gitau; Martin Paulson; Bjørn-Jostein Singstad; Karl Thomas Hjelmervik; Ola Marius Lysaker; Veralia Gabriela Sanchez

cs.CVMar 20, 2026

Multi-Stage Fine-Tuning of Pathology Foundation Models with Head-Diverse Ensembling for White Blood Cell Classification

Authors: Antony Gitau, Martin Paulson, Bjørn-Jostein Singstad, Karl Thomas Hjelmervik, Ola Marius Lysaker, Veralia Gabriela Sanchez

Organizations: University of South-Eastern Norway · Vestfold Hospital Trust

Abstract

The classification of white blood cells (WBCs) from peripheral blood smears is critical for the diagnosis of leukemia. However, automated approaches still struggle due to challenges including class imbalance, domain shift, and morphological continuum confusion, where adjacent maturation stages exhibit subtle, overlapping features. We present a multi-stage fine-tuning methodology for 13-class WBC classification in the WBCBench 2026 Challenge (ISBI 2026). Our best-performing model is a fine-tuned DINOBloom-base, on which we train multiple classifier head families (linear, cosine, and multilayer perceptron (MLP)). The cosine head performed best on the mature granulocyte boundary (Band neutrophil (BNE) F1 = 0.470), the linear head on more immature granulocyte classes (Metamyelocyte (MMY) F1 = 0.585), and the MLP head on the most immature granulocyte (Promyelocyte (PMY) F1 = 0.733), revealing class-specific specialization. Based on this specialization, we construct a head-diverse ensemble, where the MLP head acts as the primary predictor, and its predictions within the four predefined confusion pairs are replaced only when two other head families agree. We further show that cases consistently misclassified by all models are substantially enriched for probable labeling errors or inherent morphological ambiguity.

Figures & tables

Figure 1: An illustration of the morphological continuum from promyelocyte (PMY), to myelocyte (MY), metamyelocyte (MMY), band-form neutrophil (BNE) and segmented neutrophil (SNE)

Figure 2: End-to-end fine-tuning and inference simplified visual illustration. During training, separate models are obtained by fine-tuning a pathology foundation model (DINOBloom) with different classifier heads (linear, cosine, and MLP) across staged optimization. During inference, saved full DinoBloom-base and head checkpoints are combined using a head-diverse ensemble, where an MLP head acts as the primary predictor and is conditionally overridden by agreement between auxiliary heads.

Figure 3: Distributions of white blood cell types in the training set.

Figure 4: Comparison of various of our models’ performances on the validation and test sets

Figure 5: Comparison of classification head performance for the boundary subsets.

Figure 6: Evaluation confusion matrix of the final hybrid. Highlighted regions indicate recurring confusions relevant to the conservative override rule.

Disagreement	$n$	Label Error	Model Error	Ambig.
Train	1,457	803 (55.1%)	484 (33.2%)	170 (11.7%)
Validation	332	172 (51.8%)	140 (42.2%)	20 (6.0%)
Combined	1,789	975 (54.5%)	624 (34.9%)	190 (10.6%)
Agreement	$n$	Label Error	Ambig.	Confirmed Correct
Train	1,954	156 (8.0%)	125 (6.4%)	1,673 (85.6%)

Table 1: Summary of expert label review outcomes.

Figure 7: Directional analysis of reviewed disagreement cases across training and validation sets. Rows denote assigned labels and columns denote model top-1 predictions. Each cell shows label errors / reviewed disagreements for that pair, and color indicates the label-error rate..

Explore similar work

Apr 25, 2026

A Hierarchical Ensemble Inference Pipeline for Robust White Blood Cell Classification Under Domain Shifts

Automated white blood cell (WBC) classification is essential for scalable leukaemia screening. However, real-world deployment is challenged by domain shifts caused by staining protocols, scanner characteristics, and inter-laboratory variability, which often degrade model performance. The White Blood Cell Classification Challenge (WBCBench) at ISBI 2026 aims to advance robust WBC recognition, with a focus on accurately identifying blast cells and other clinically critical rare subtypes. We propose a memory-augmented, hierarchical ensemble pipeline for WBC classification under domain shifts, leveraging a feature bank and a DinoBloom backbone fine-tuned with LoRA. Our three-stage inference hierarchy combines k-nearest neighbors (kNN) retrieval at each level, reducing over-reliance on any single decision. Evaluated on the WBCBench dataset, our method ranks within the top ten by macro F1-score in the final testing phase.

Ruyi Dai, Tingkwong Ng, Hao Chen

Department of Computer Science and Engineering, The Hong Kong University of Science and Technology, Hong Kong, China · Department of Chemical and Biological Engineering, The Hong Kong University of Science and Technology, Hong Kong, China · Division of Life Science, Hong Kong University of Science and Technology, Hong Kong, China +2

Hierarchical Classification Ensemble Learning Domain Generalization Medical Image Classification

Ask PDF

Apr 12, 2026

WBCBench 2026: A Challenge for Robust White Blood Cell Classification Under Class Imbalance

We present WBCBench 2026, an ISBI challenge and benchmark for automated WBC classification designed to stress-test algorithms under three key difficulties: (i) severe class imbalance across 13 morphologically fine-grained WBC classes, (ii) strict patient-level separation between training, validation and test sets, and (iii) synthetic scanner- and setting-induced domain shift via controlled noise, blur and illumination perturbations. All images are single-site microscopic blood smear acquisitions with standardised staining and expert hematopathologist annotations. This paper reviews the challenge and summarises the proposed solutions and final outcomes. The benchmark is organised into two phases. Phase 1 provides a pristine training set. Phase 2 introduces degraded images with split-specific severity distributions for train, validation and test, emulating a realistic shift between development and deployment conditions. We specify a standardised submission schema, open-source evaluator, and macro-averaged F1 score as the primary ranking metric.

Xin Tian, Xudong Ma, Tianqi Yang +4

University of Oxford, Oxford, United Kingdom · University College London, London, United Kingdom · University of Bristol, Bristol, United Kingdom +1

Medical Image Classification Medical Image Benchmarks Distribution Shift Robustness Fine-Grained Image Classification Class-Imbalanced Learning

Ask PDF

Mar 21, 2026

Ensemble of Small Classifiers For Imbalanced White Blood Cell Classification

Automating white blood cell classification for diagnosis of leukaemia is a promising alternative to time-consuming and resource-intensive examination of cells by expert pathologists. However, designing robust algorithms for classification of rare cell types remains challenging due to variations in staining, scanning and inter-patient heterogeneity. We propose a lightweight ensemble approach for classification of cells during Haematopoiesis, with a focus on the biology of Granulopoiesis, Monocytopoiesis and Lymphopoiesis. Through dataset expansion to alleviate some class imbalance, we demonstrate that a simple ensemble of lightweight pretrained SwinV2-Tiny, DinoBloom-Small and ConvNeXT-V2-Tiny models achieves excellent performance on this challenging dataset. We train 3 instantiations of each architecture in a stratified 3-fold cross-validation framework; for an input image, we forward-pass through all 9 models and aggregate through logit averaging. We further reason on the weaknesses of our model in confusing similar-looking myelocytes in granulopoiesis and lymphocytes in lymphopoiesis. Code: https://gitlab.com/siddharthsrivastava/wbc-bench-2026.

Siddharth Srivastava, Adam Smith, Scott Brooks +2

Department of Computer Science, University of Warwick · Warwick Medical School, University of Warwick · Intelligent Imaging Innovations Ltd

Medical Image Classification

Ask PDF