Ensemble of Small Classifiers For Imbalanced White Blood Cell Classification

2026-03-21

AI Summary

This work addresses the challenge of classifying rare white blood cell types in leukemia diagnosis, where staining inconsistencies, scanning variations, and patient heterogeneity—particularly the class imbalance and morphological ambiguity among granulocytic, monocytic, and lymphoid lineages—hinder accurate identification. To tackle this, we propose a lightweight multi-model ensemble strategy that mitigates class imbalance through data augmentation and integrates three efficient vision backbones: SwinV2-Tiny, DinoBloom-Small, and ConvNeXt-V2-Tiny. Nine models are trained under stratified three-fold cross-validation, with final predictions aggregated via logit averaging. Our approach substantially improves recognition accuracy for rare cell types and provides insight into the model’s confusion mechanisms between morphologically similar cells, such as promyelocytes and lymphocytes.

Technology Category

• Machine Learning: Ensemble Methods
• Computer Vision: Multi-modal Vision
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models

Application Category

• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

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.

Problem

• imbalanced classification
• white blood cell classification
• rare cell types
• inter-patient heterogeneity
• morphological similarity

Innovation

• ensemble learning
• class imbalance
• white blood cell classification
• lightweight models
• logit averaging