White blood cell classification across heterogeneous image datasets remains challenging due to variations in staining protocols, scanning devices, resolution, and class distribution. In this work, we propose a two-stage framework for robust cross-dataset classification. YOLOv11 is first applied for white blood cell localization and extraction, followed by a Swin Transformer for fine-grained classification. Experimental results on multi-source datasets demonstrate that the proposed method outperforms ResNet50, VGG16-BN, and EfficientNet-B0. The best 9-class no-crop setting achieves 0.9709 macro-F1 on mixed-source testing, while maintaining 0.8188 macro-F1 on an unseen external test set.

Index Terms: WBC classification, cross-dataset generalization, YOLOv11, Swin Transformer, class imbalance

Presented at the Work-in-Progress poster session "Image Classification and Supervised Representation Learning", IEEE ICIP 2026, Tampere Hall (Winter Garden), Tampere, Finland, 15 September 2026, Paper #3846.

Affiliations: Che-Wei Chang and Chien-Yu Chen, Department of Biomechatronics Engineering, National Taiwan University; Kuo-Jui Sun, MacKay Memorial Hospital; Cheng-Hong Tsai, National Taiwan University Hospital.