Wonjun Lee
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Wonjun Lee has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Sindy Kam Yan Tang
Department of Mechanical Engineering, Stanford University, Stanford, USA
2 shared publications - Byoungkwon Yoon
Department of Mechanical Engineering, Seoul National University, Seoul, Republic of Korea
1 shared publication Chien-Yi Chang
Department of Electrical Engineering, Stanford University, Stanford, CA 94305, USA
1 shared publication- Jihoon Ko
Department of BioNano Technology Gachon University Seongnam‐si Gyeonggi‐do Republic of Korea
1 shared publication - Jungseub Lee
Department of Mechanical Engineering, Seoul National University, Seoul, Republic of Korea
1 shared publication - Leon Hannes Kloker
Biohub
1 shared publication
Latest Publications
- Open Access
Code-Multiplexed Multi-Frequency Impedance Cytometry with a Unified Deep-Unfolding Network
Impedance flow cytometry (IFC) is a label-free, single-cell measurement technique that captures biophysical properties beyond traditional biochemical markers. Code-multiplexing allows parallelization of IFC with simple hardware but requires advanced signal processing algorithm to resolve overlaps in signals originating from different channels. Existing methods, however, rely on multiple task-specific networks with template-based linear fitting, which loses accuracy under nonlinear or unstable conditions common in microfluidic experiments. Prior studies have also been restricted to demultiplexing single-frequency impedance measurements. To this end, we develop a unified deep-unfolding network for analyzing code multiplexed, multi frequency IFC data. We unfold successive-interference cancellation (SIC) algorithm into a deep-learning network that encodes the structural priors of SIC, and enable a single multitask network to resolve overlaps while sharing information across tasks. To mitigate nonlinear signal stretching and amplification, our network recognizes events by predicting bit level intensity and duration. For multi-frequency impedance profiling, we apply least-squares fitting to the predicted single-frequency real-impedance trace to map the real and imaginary impedance traces at other frequencies. On the cell-bead mixture evaluation dataset, our pipeline resolves overlaps from singlets to triplets reliably, reconstructs impedance-intensity distributions accurately, and enables multi-frequency impedance profiling. As a demonstration of principle, we use our pipeline to perform label free quantification of basophil activation from code-multiplexed, multi-frequency IFC measurements. Consistent with previous study, impedance opacity correlates well with activation levels measured by fluorescence flow cytometry. In summary, our study demonstrates the feasibility and utility of a deep-unfolding network that extends code-multiplexed, multi-frequency IFC to label-free single-cell functional assays.
Open AccessCode-Multiplexed Multi-Frequency Impedance Cytometry with a Unified Deep-Unfolding Network
Impedance flow cytometry (IFC) is a label-free, single-cell measurement technique that captures biophysical properties beyond traditional biochemical markers. Code-multiplexing allows parallelization of IFC with simple hardware but requires advanced signal processing algorithm to resolve overlaps in signals originating from different channels. Existing methods, however, rely on multiple task-specific networks with template-based linear fitting, which loses accuracy under nonlinear or unstable conditions common in microfluidic experiments. Prior studies have also been restricted to demultiplexing single-frequency impedance measurements. To this end, we develop a unified deep-unfolding network for analyzing code multiplexed, multi frequency IFC data. We unfold successive-interference cancellation (SIC) algorithm into a deep-learning network that encodes the structural priors of SIC, and enable a single multitask network to resolve overlaps while sharing information across tasks. To mitigate nonlinear signal stretching and amplification, our network recognizes events by predicting bit level intensity and duration. For multi-frequency impedance profiling, we apply least-squares fitting to the predicted single-frequency real-impedance trace to map the real and imaginary impedance traces at other frequencies. On the cell-bead mixture evaluation dataset, our pipeline resolves overlaps from singlets to triplets reliably, reconstructs impedance-intensity distributions accurately, and enables multi-frequency impedance profiling. As a demonstration of principle, we use our pipeline to perform label free quantification of basophil activation from code-multiplexed, multi-frequency IFC measurements. Consistent with previous study, impedance opacity correlates well with activation levels measured by fluorescence flow cytometry. In summary, our study demonstrates the feasibility and utility of a deep-unfolding network that extends code-multiplexed, multi-frequency IFC to label-free single-cell functional assays.
- 25 September 2025
- Published by Elsevier in Sensors and Actuators B: Chemical
- Vol. 447 (Pt 2)
Label-free, machine-learning assisted impedance assessment of basophil activation
- 25 September 2025
- Published by Elsevier in Sensors and Actuators B: Chemical
- Vol. 447 (Pt 2)
Label-free, machine-learning assisted impedance assessment of basophil activation
- 31 July 2023
- Published by Springer Nature in BioChip Journal
- Vol. 17 (3) , 357-368
Machine Learning-Aided Three-Dimensional Morphological Quantification of Angiogenic Vasculature in the Multiculture Microfluidic Platform
- 31 July 2023
- Published by Springer Nature in BioChip Journal
- Vol. 17 (3) , 357-368
Machine Learning-Aided Three-Dimensional Morphological Quantification of Angiogenic Vasculature in the Multiculture Microfluidic Platform