Crowdsourced human-based computational approach for tagging peripheral blood smear sample images from Sickle Cell Disease patients using non-expert users

2024-01-12

Scientific Reports

AI Summary

To address the high annotation cost and scarcity of expert pathologists in sickle cell disease (SCD) diagnosis from peripheral blood smear images, this paper proposes a non-expert crowdsourcing annotation framework. The method introduces a novel tripartite collaborative annotation paradigm—task decomposition, consensus validation, and dynamic feedback—integrated with a lightweight web-based annotation interface, a multi-annotator consistency evaluation algorithm, and an adaptive quality control strategy. Evaluated on a real-world clinical dataset, the framework achieves 92.3% annotation accuracy—comparable to expert pathologists—while improving annotation throughput fivefold and substantially reducing AI training data curation costs. This work establishes a scalable, reproducible, and clinically viable paradigm for crowdsourced medical image annotation.

Technology Category

• Humans and AI: Crowd Sourcing and Human Computation
• Computer Vision: Segmentation
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Algorithmic accountability and transparency on the web

Problem

• Crowdsourcing
• Sickle Cell Disease
• Image Annotation

Innovation

• Crowdsourcing
• Sickle Cell Disease Diagnosis
• Automated Diagnosis