AI-based portable liquid phase leukocyte detection system.

BACKGROUND: White blood cells (WBCs) are vital components of the immune system, and accurate detection and classification are essential for diagnosing and monitoring various diseases. Traditional microscopic detection methods require expensive equipment and skilled personnel, limiting their use in primary care and resource-constrained environments. To address these challenges, we propose a portable, low-cost, and intelligent system for WBC detection in liquid-phase samples, aiming to simplify operation and improve accessibility in point-of-care testing (POCT) settings. RESULTS: The developed system integrates a compact hardware platform with rotating multi-view imaging and artificial intelligence-based analysis. Multi-focal length image fusion mitigates defocus issues, and attention mechanisms enhance key feature extraction under suboptimal imaging conditions. The detection model achieves error rates below 1 % for WBC counting and ∼2 % for classification. Clinical validation using 80 samples showed strong consistency with reference methods, yielding correlation coefficients of 0.99469 (lymphocytes), 0.98436 (neutrophils), and 0.98625 (monocytes), with an approximate deviation of 4 %. The system operates with simple loading and initiation steps, ensuring ease of use in decentralized medical settings. SIGNIFICANCE: This work demonstrates a practical solution for automated WBC detection in primary care environments, offering low cost, high accuracy, and user-friendly operation. It has strong potential to enhance the efficiency, accessibility, and standardization of hematological diagnostics in POCT applications.