Development and Validation of an Unsupervised Feature Learning System for Leukocyte Characterization and Classification: A Multi-Hospital Study

Hong Yan,Xuanyu Mao,Xu Yang,Yongquan Xia,Chengbin Wang,Junjun Wang,Rui Xia,Xuejing Xu,Zhiqiang Wang,Zhiyang Li,Xie Zhao,Yan Li,Guoye Liu,Li He,Zhongyu Wang,Zhiqiong Wang,Zhiqiang Li,Weidong Cai,Han Shen,Hang Chang

DOI: https://doi.org/10.1007/s11263-021-01449-9

IF: 13.369

2021-03-27

International Journal of Computer Vision

Abstract:The characterization and classification of white blood cells (WBC) are critical for the diagnosis of anemia, leukemia, and many other hematologic diseases. We developed WBC-Profiler, an unsupervised feature learning system for quantitative analysis of leukocytes. We demonstrate, through independent validation, that WBC-Profiler enables automatic extraction of complex and robust signatures from microscopic images without human-intervention and, thereafter, effective construction of interpretable leukocyte profiles, which decouples large scale complex leukocyte characterization from limitations in both human-based feature engineering/optimization and the end-to-end solutions provided by many modern deep neural networks. Further evaluation in a real-world clinical setting confirms that, compared with 23 clinicians from 8 hospitals (class-average-sensitivity, 0.798; class-average-specificity, 0.963; cell-average-timecost: 3.158 s), WBC-Profiler performs with significantly improved accuracy and speed (class-average-sensitivity, 0.890; class-average-specificity, 0.980; cell-average-timecost: 0.375 s). Our findings suggest that WBC-Profiler has the potential clinical implications.

computer science, artificial intelligence

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to address the problem of white blood cell (WBC) characterization and classification, particularly the need for high-precision automatic recognition of white blood cells in clinical laboratories. Specifically: 1. **Importance of WBC characterization and classification**: - WBC characterization and classification are crucial for diagnosing various blood diseases such as anemia and leukemia. - Existing methods rely on manual annotation, which is not only time-consuming but also lacks experienced pathologists in many clinical laboratories. 2. **Limitations of existing technologies**: - Early attempts at commercial systems failed due to high costs and low accuracy. - Recent methods based on deep neural networks have improved classification accuracy, but most systems focus on end-to-end solutions, neglecting detailed explanations of WBC features. 3. **Proposed new method**: - The paper develops the WBC-Profiler system, an unsupervised feature learning system that can automatically extract complex and robust feature signatures from microscopic images and construct interpretable WBC characterizations. - This system not only improves classification accuracy and speed but also provides detailed feature explanations, aiding in WBC characterization analysis in clinical applications. ### Specific research content: - **Dataset**: The study uses datasets from multiple hospitals in China, including data from Gulou Hospital and 301 Hospital. - **Experimental validation**: Through multi-center clinical validation, it was demonstrated that WBC-Profiler significantly outperformed 23 clinical doctors from 8 hospitals in actual clinical environments. - **Performance metrics**: WBC-Profiler excelled in classification accuracy (class-average-sensitivity), specificity (class-average-specificity), and processing speed. Through this research, the paper showcases the potential of WBC-Profiler in WBC characterization and classification tasks and emphasizes its potential application value in clinical practice.

Development and Validation of an Unsupervised Feature Learning System for Leukocyte Characterization and Classification: A Multi-Hospital Study

Deep transfer learning technique to detect white blood cell classification in regular clinical practice using histopathological images

Development and Evaluation of a Leukemia Diagnosis System Using Deep Learning in Real Clinical Scenarios

Hybrid adversarial-discriminative network for leukocyte classification in leukemia

Label-free white blood cells classification using a deep feature fusion neural network

Efficient leukocytes detection and classification in microscopic blood images using convolutional neural network coupled with a dual attention network

White Blood Cell Classification Using Texture and RGB Features of Oversampled Microscopic Images

Automatic Classification of Leukocytes Using Deep Neural Network

Accurate Leukocyte Detection Based on Deformable-DETR and Multi-Level Feature Fusion for Aiding Diagnosis of Blood Diseases

High-Efficiency Classification of White Blood Cells Based on Object Detection

AI-Based Portable White Blood Cells Classification and Counting System in POCT

Segmentation, feature extraction and classification of leukocytes leveraging neural networks, a comparative study

Automated White Blood Cell Disease Recognition Using Lightweight Deep Learning

Development and application of automatic recognition system for leukocyte images

Attention-Aware Residual Network Based Manifold Learning for White Blood Cells Classification.

Automatic classification of white blood cells using deep features based convolutional neural network

Accurate label-free 3-part leukocyte recognition with single cell lens-free imaging flow cytometry

Developing and Preliminary Validating an Automatic Cell Classification System for Bone Marrow Smears: a Pilot Study

Advanced Convolutional Neural Networks for Precise White Blood Cell Subtype Classification in Medical Diagnostics

Deep learning for few-shot white blood cell image classification and feature learning