Digital Imaging for Blood Diseases

Trust, but Verify: Epistemic Uncertainty-Gated Pseudo-Labeling for Safer Semi-Supervised Malaria Screening

This paper introduces an uncertainty-gated pseudo-labeling method for semi-supervised malaria screening, reducing silent failures without sacrificing accuracy.

Quantifying performance inflation from hidden sample dependence in biomedical image classification benchmarks

The paper quantifies how increasing train-test sample relatedness inflates classification performance in biomedical image benchmarks using a microscopy dataset.

Decoupling Detection and Classification to Improve Morphological Phenotype Analysis of Sickle Red Blood Cells in Full-Scope Microscopy

This paper improves sickle cell disease morphology analysis by combining a YOLO detector with a DenseNet classifier for accurate detection and classification of red blood cell morphologies in full-scope microscopy images.

BUsmear: A Low-Cost 3D-Printed Automated Device for Blood Smear Preparation

BUsmear is a low-cost 3D-printed device that prepares uniform blood smears, improving reproducibility and morphology compared to manual methods.

Interpretable Peripheral Blood Cell Classification via Vision-Language Concept Bottleneck and Soft Decision Tree

The paper introduces an interpretable pipeline for peripheral blood cell classification using a vision-language model and a soft decision tree, achieving high accuracy with traceable decision logic.

Plasmo3Net: A Convolutional Neural Network-Based Algorithm for Detecting Malaria Parasites in Thin Blood Smear Images

This paper develops Plasmo3Net, a CNN-based algorithm for detecting malaria parasites in thin blood smears, achieving high accuracy and precision.

Interpretable morphology mapping of peripheral blood leukocytes using annotation-efficient artificial intelligence

This paper develops an interpretable AI framework for leukocyte morphology classification in peripheral blood smears, achieving high accuracy with reduced annotation needs and providing clinical tools for integration.

Majority-LCL: Towards Malaria Cell Detection using Label Contrastive Learning and Majority voting Ensembling

This paper introduces a malaria cell detection method combining label contrastive learning and majority voting ensembling to improve diagnostic accuracy using microscopic images.

BoMBR: An Annotated Bone Marrow Biopsy Dataset for Segmentation of Reticulin Fibers

The paper introduces BoMBR, a dataset of 201 annotated bone marrow biopsy images for reticulin fiber segmentation, and demonstrates a deep learning model achieving high accuracy in reticulin quantification and marrow fibrosis grading.

BaMBo: An Annotated Bone Marrow Biopsy Dataset for Segmentation Task

The paper introduces BaMBo, a bone marrow biopsy dataset for automated cellularity calculation, and presents a U-Net model achieving 83.1% Dice score and 5.9% cellularity prediction error.

Automatic Classification of Circulating Blood Cell Clusters based on Multi-channel Flow Cytometry Imaging

This study introduces a framework for automatically classifying circulating blood cell clusters using flow cytometry images, achieving over 95% accuracy in cluster classification and phenotype identification.

Cross-Species Morphology Learning Enables Nucleic Acid-Independent Detection of Live Mutant Blood Cells

This paper introduces a computer vision platform that detects live mutant blood cells with specific mutations using morphology recognition, without requiring nucleic acid sequencing.

Leveraging Sparse Annotations for Leukemia Diagnosis on the Large Leukemia Dataset

This paper introduces a large leukemia dataset and methods for WBC detection with attribute analysis using sparse annotations to improve diagnostic accuracy and reduce annotation burden.

Application of machine learning in a rodent malaria model for rapid, accurate, and consistent parasite counts

The paper develops a machine learning tool for automated parasite counting in rodent malaria models, achieving high accuracy and consistency across labs.

DeepHeme: A generalizable, bone marrow classifierwith hematopathologist-level performance

DeepHeme is a convolutional neural network trained on bone marrow single-cell images that achieves high accuracy in classifying hematologic cell types, outperforming human experts.

Real-time detection of acute lymphoblastic leukemia cells using deep learning

This study uses deep learning models to detect acute lymphoblastic leukemia cells in microscopic images, achieving 97.2% accuracy and 80 frames per second processing speed.

Data augmentation for imbalanced blood cell image classification

This paper proposes a method to improve classification of imbalanced blood cell images using data augmentation techniques and a minority class sampling strategy, achieving better performance on minority classes.

Detection of live breast cancer cells in brightfield microscopy images containing white blood cells by image analysis and deep learning

This paper develops deep learning methods to detect unstained breast cancer cells in brightfield images containing white blood cells, achieving high sensitivity and specificity.

The Evaluation of Acute Myeloid Leukaemia (AML) Blood Cell Detection Models Using Different YOLO Approaches

This paper evaluates YOLO models for detecting AML blast cells in blood images, finding YOLOv3 performs best with over 92% precision and sensitivity.

Label-free assessment of red blood cell storage lesions by deep learning

This study uses deep learning and label-free imaging flow cytometry to classify red blood cell morphologies during storage, achieving 90% accuracy and predicting expiration dates based on morphological changes.

Classification and Generation of MicroscopyImages with Plasmodium Falciparum via ArtificialNeural Networks

This paper develops a deep learning model for malaria parasite detection and synthetic image generation in microscopy images, achieving high classification accuracy and creating realistic images indistinguishable from real data.

Classification of human white blood cells using machine learning for stain-free imaging flow cytometry

This study compares classical machine learning and deep learning for classifying white blood cells using stain-free imaging flow cytometry, achieving balanced accuracies of 0.776 and 0.697 for white blood cells and 0.866 and 0.867 for eosinophils.

WBC-Profiler: an Unsupervised Feature Learning System for Leukocytes Characterization and Classification

WBC-Profiler automatically extracts complex signatures from microscopic images to construct leukocyte profiles without human intervention.