Related Experiment Video
Updated: Dec 9, 2025
07:27
Quantifying Microorganisms at Low Concentrations Using Digital Holographic Microscopy DHM
Published on: November 1, 2017
10.8K
Deep learning-based cell identification and disease diagnosis using spatio-temporal cellular dynamics in compact
Timothy O'Connor1, Arun Anand2, Biree Andemariam3
1Biomedical Engineering Department, University of Connecticut, Storrs, Connecticut 06269, USA.
Biomedical Optics Express
|September 14, 2020
Summary
This study introduces a deep learning method using digital holographic microscopy to identify cells and diagnose diseases by analyzing cell behavior over time. The approach accurately distinguishes cell types and detects sickle cell disease.
Area of Science:
- Biomedical Engineering
- Computational Biology
- Microscopy
Background:
- Accurate cell identification and disease diagnosis are crucial in healthcare.
- Traditional methods often struggle with morphologically similar cells or subtle disease indicators.
- Digital holographic microscopy offers high-resolution, label-free imaging of live cells.
Purpose of the Study:
- To develop and validate a deep learning strategy for cell identification and disease diagnosis.
- To leverage spatio-temporal information from digital holographic microscopy for enhanced classification.
- To demonstrate the system's efficacy in distinguishing similar cell types and detecting human red blood cell diseases.
Main Methods:
- Utilized shearing digital holographic microscopy for video-rate data acquisition of live cells.
- Extracted features from reconstructed phase profiles of segmented cells.
- Employed a recurrent bi-directional long short-term memory (Bi-LSTM) network for spatio-temporal analysis and classification.
- Applied the method to differentiate cow and horse red blood cells and diagnose sickle cell disease in human red blood cells.
Main Results:
- Successfully identified and classified cells based on their dynamic behavior.
- Achieved improved diagnostic performance compared to conventional machine learning methods for sickle cell disease detection.
- Demonstrated accurate classification at both cellular and patient levels.
Conclusions:
- Deep learning combined with digital holographic microscopy provides a powerful tool for cell identification and disease diagnosis.
- The spatio-temporal analysis of cell behavior offers novel insights for biomedical applications.
- This represents a pioneering application of deep learning for cell analysis using digital holographic microscopy.
