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

• URL: http://arxiv.org/abs/2603.20383v1
• Date: Fri, 20 Mar 2026 18:00:59 GMT
• Title: Multi-Stage Fine-Tuning of Pathology Foundation Models with Head-Diverse Ensembling for White Blood Cell Classification
• Abstract summary: We present a multi-stage fine-tuning methodology for 13-class WBC classification.<n>Our best-performing model is a fine-tuned DINOBloom-base.<n>Cases consistently misclassified by all models are substantially enriched for probable labeling errors or inherent morphological ambiguity.
• Score: 0.08209843760716957
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• 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.

Related papers

• A Hierarchical Ensemble Inference Pipeline for Robust White Blood Cell Classification Under Domain Shifts [5.200916021795385]
  We propose a memory-augmented, hierarchical ensemble pipeline for WBC classification under domain shifts.<n>Our method ranks within the top ten by macro F1-score in the final testing phase.
  arXiv  Detail & Related papers  (2026-04-25T12:25:46Z)
• Robust White Blood Cell Classification with Stain-Normalized Decoupled Learning and Ensembling [6.750164263871006]
  White blood cell (WBC) classification is fundamental for hematology applications such as infection assessment, leukemia screening, and treatment monitoring.<n>We propose a stain-normalized, decoupled training framework that learns transferable representations using instance-balanced sampling.<n>In inference stage, we further enhance robustness by ensembling various trained backbones with test-time augmentation.
  arXiv  Detail & Related papers  (2026-03-02T15:31:25Z)
• MeCaMIL: Causality-Aware Multiple Instance Learning for Fair and Interpretable Whole Slide Image Diagnosis [40.3028468133626]
  Multiple instance learning (MIL) has emerged as the dominant paradigm for whole slide image (WSI) analysis in computational pathology.<n>textbfMeCaMIL, a causality-aware MIL framework, explicitly models demographic confounders through structured causal graphs.<n>MeCaMIL achieves superior fairness -- demographic disparity variance drops by over 65% relative reduction on average across attributes.
  arXiv  Detail & Related papers  (2025-11-14T06:47:21Z)
• Deep Learning Approaches for Blood Disease Diagnosis Across Hematopoietic Lineages [0.0]
  We present a foundation modeling framework that leverages deep learning to uncover latent genetic signatures across the hematopoietic hierarchy.<n>Our approach trains a fully connected autoencoder on multipotent progenitor cells, reducing over 20,000 gene features to a 256-dimensional latent space.<n>We validate the quality of these embeddings by training feed-forward, transformer, and graph convolutional architectures for blood disease diagnosis tasks.<n>Our models achieve greater than 95% accuracy for multi-class classification, and in the zero-shot setting, we achieve greater than 0.7 F1-score on the binary classification task.
  arXiv  Detail & Related papers  (2025-03-25T20:11:10Z)
• Pathological Prior-Guided Multiple Instance Learning For Mitigating Catastrophic Forgetting in Breast Cancer Whole Slide Image Classification [50.899861205016265]
  We propose a new framework PaGMIL to mitigate catastrophic forgetting in breast cancer WSI classification.<n>Our framework introduces two key components into the common MIL model architecture.<n>We evaluate the continual learning performance of PaGMIL across several public breast cancer datasets.
  arXiv  Detail & Related papers  (2025-03-08T04:51:58Z)
• MMIL: A novel algorithm for disease associated cell type discovery [58.044870442206914]
  Single-cell datasets often lack individual cell labels, making it challenging to identify cells associated with disease. We introduce Mixture Modeling for Multiple Learning Instance (MMIL), an expectation method that enables the training and calibration of cell-level classifiers.
  arXiv  Detail & Related papers  (2024-06-12T15:22:56Z)
• Neural Cellular Automata for Lightweight, Robust and Explainable Classification of White Blood Cell Images [40.347953893940044]
  We introduce a novel approach for white blood cell classification based on neural cellular automata (NCA) Our NCA-based method is significantly smaller in terms of parameters and exhibits robustness to domain shifts. Our results demonstrate that NCA can be used for image classification, and they address key challenges of conventional methods.
  arXiv  Detail & Related papers  (2024-04-08T14:59:53Z)
• A Study of Age and Sex Bias in Multiple Instance Learning based Classification of Acute Myeloid Leukemia Subtypes [44.077241051884926]
  We train multiple MIL models using different levels of sex imbalance in the training set and excluding certain age groups. We find a significant effect of sex and age bias on the performance of the model for AML subtype classification.
  arXiv  Detail & Related papers  (2023-08-24T09:32:46Z)
• Imbalanced Domain Generalization for Robust Single Cell Classification in Hematological Cytomorphology [3.7007225479462402]
  We train a robust CNN for WBC classification by addressing cross-domain data imbalance and domain shifts. Our approach achieves the best F1 macro score compared to other existing methods.
  arXiv  Detail & Related papers  (2023-03-14T10:20:31Z)
• Deep learning-based Subtyping of Atypical and Normal Mitoses using a Hierarchical Anchor-Free Object Detector [0.802219018904343]
  Atypical mitotic figures (MF) can be identified morphologically as having segregation abnormalities of the chromatids. In this work, we perform, for the first time, automatic subtyping of mitotic figures into normal and atypical categories. We set up a state-of-the-art object detection pipeline extending the anchor-free FCOS approach with a gated hierarchical subclassification branch.
  arXiv  Detail & Related papers  (2022-12-12T13:57:38Z)
• Medulloblastoma Tumor Classification using Deep Transfer Learning with Multi-Scale EfficientNets [63.62764375279861]
  We propose an end-to-end MB tumor classification and explore transfer learning with various input sizes and matching network dimensions. Using a data set with 161 cases, we demonstrate that pre-trained EfficientNets with larger input resolutions lead to significant performance improvements.
  arXiv  Detail & Related papers  (2021-09-10T13:07:11Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.