Exploring Few-Shot Object Detection on Blood Smear Images: A Case Study of Leukocytes and Schistocytes

2025-03-21

AI Summary

This study addresses few-shot detection of white blood cells and schistocytes in blood smear images, tackling performance degradation caused by clinical annotation scarcity and domain shift. We propose DE-ViT—the first vision transformer specifically designed for few-shot medical object detection—and systematically benchmark it against Faster R-CNN (with ResNet-50/X101 backbones), conventional few-shot learning, and transfer learning approaches. Experiments on the Raabin-WBC and SC-IDB datasets reveal that Faster R-CNN with ResNet-X101 achieves the highest detection accuracy; while DE-ViT introduces architectural novelty, its performance is severely hampered by cross-domain distribution shift. Crucially, this work provides the first empirical evidence that domain shift—not merely label scarcity—is the dominant bottleneck in medical few-shot detection. These findings establish a critical empirical baseline and offer theoretical guidance for future method development in domain-adaptive few-shot medical imaging analysis.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Security and Privacy: Data transparency and provenance
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

The detection of blood disorders often hinges upon the quantification of specific blood cell types. Variations in cell counts may indicate the presence of pathological conditions. Thus, the significance of developing precise automatic systems for blood cell enumeration is underscored. The investigation focuses on a novel approach termed DE-ViT. This methodology is employed in a Few-Shot paradigm, wherein training relies on a limited number of images. Two distinct datasets are utilised for experimental purposes: the Raabin-WBC dataset for Leukocyte detection and a local dataset for Schistocyte identification. In addition to the DE-ViT model, two baseline models, Faster R-CNN 50 and Faster R-CNN X 101, are employed, with their outcomes being compared against those of the proposed model. While DE-ViT has demonstrated state-of-the-art performance on the COCO and LVIS datasets, both baseline models surpassed its performance on the Raabin-WBC dataset. Moreover, only Faster R-CNN X 101 yielded satisfactory results on the SC-IDB. The observed disparities in performance may possibly be attributed to domain shift phenomena.

Problem

• Develop precise automatic systems for blood cell detection
• Explore Few-Shot object detection on blood smear images
• Compare DE-ViT with baseline models for leukocyte and schistocyte identification

Innovation

• DE-ViT model for Few-Shot detection
• Utilizes Raabin-WBC and local datasets
• Compares with Faster R-CNN baselines