More Like this

1. Comparing machine learning and deep learning regression frameworks for accurate prediction of dielectrophoretic force

   https://doi.org/10.1038/s41598-022-16114-5  

   Ajala, Sunday; Muraleedharan Jalajamony, Harikrishnan; Nair, Midhun; Marimuthu, Pradeep; Fernandez, Renny Edwin (July 2022, Scientific Reports)

   Abstract An intelligent sensing framework using Machine Learning (ML) and Deep Learning (DL) architectures to precisely quantify dielectrophoretic force invoked on microparticles in a textile electrode-based DEP sensing device is reported. The prediction accuracy and generalization ability of the framework was validated using experimental results. Images of pearl chain alignment at varying input voltages were used to build deep regression models using modified ML and CNN architectures that can correlate pearl chain alignment patterns of Saccharomyces cerevisiae(yeast) cells and polystyrene microbeads to DEP force. Various ML models such as K-Nearest Neighbor, Support Vector Machine, Random Forest, Neural Networks, and Linear Regression along with DL models such as Convolutional Neural Network (CNN) architectures of AlexNet, ResNet-50, MobileNetV2, and GoogLeNet have been analyzed in order to build an effective regression framework to estimate the force induced on yeast cells and microbeads. The efficiencies of the models were evaluated using Mean Absolute Error, Mean Absolute Relative, Mean Squared Error, R-squared, and Root Mean Square Error (RMSE) as evaluation metrics. ResNet-50 with RMSPROP gave the best performance, with a validation RMSE of 0.0918 on yeast cells while AlexNet with ADAM optimizer gave the best performance, with a validation RMSE of 0.1745 on microbeads. This provides a baseline for further studies in the application of deep learning in DEP aided Lab-on-Chip devices. 

   more » « less
2. Dielectrophoretic trapping of particles flowing through an array of conductive cylinders

   https://doi.org/10.1063/5.0305604  

   Mondal, Tonoy K; Williams, Stuart J (December 2025, Physics of Fluids)

   Dielectrophoresis (DEP) is a label-free electrokinetic method for selectively trapping polarizable particles using non-uniform electric fields. While co-planar electrode systems are common, their inherent DEP force distribution limits throughput. This study presents a computationally efficient framework for modeling two-dimensional DEP-based particle trapping in ordered arrays of conductive cylinders. These cylinders are modeled at a range of sizes, from micrometers to nanometers, to represent microfluidic systems consisting of conductive pillars, nanofibers, etc. Analytical solutions for fluid flow and electric potential were derived using eigenfunction expansions and collocation, then used in a particle tracking model that includes hydrodynamic drag, Brownian motion, and multipolar DEP forces. Although focused on conductive arrays, this framework is extensible to other configurations. This work provides a foundation for future work in the design of high-throughput DEP systems. Both dimensionless and dimensional analyses were performed across a wide range of particle sizes (30 nm to 3 μm), voltages (10 mV to 100 V), and array geometries. No specific optimal cylinder size was found; instead, optimal performance arises from a balance between DEP force distribution and flow through the cylinder array gap. Diamond-oriented arrays exhibited enhanced trapping under moderate dielectrophoretic velocity-to-fluid velocity ratios (up to 39% greater), while square arrays performed better under low-field and large-cylinder conditions (up to 40% greater). 

   more » « less
3. Three‐dimensional lab‐on‐a‐foil device for dielectrophoretic separation of cancer cells

   https://doi.org/10.1002/elps.202200287  

   Wu, Mengren; Gao, Yuan; Luan, Qiyue; Papautsky, Ian; Chen, Xiaolin; Xu, Jie (April 2023, ELECTROPHORESIS)

   Abstract A simple, low‐cost, three‐dimensional (3D) lab‐on‐a‐foil microfluidic device for dielectrophoretic separation of circulating tumor cells (CTCs) is designed and constructed. Disposable thin films are cut by xurography and microelectrode array are made with rapid inkjet printing. The multilayer device design allows the studying of spatial movements of CTCs and red blood cells (RBCs) under dielectrophoresis (DEP). A numerical simulation was performed to find the optimum driving frequency of RBCs and the crossover frequency for CTCs. At the optimum frequency, RBCs were lifted 120 µm inz‐axis direction by DEP force, and CTCs were not affected due to negligible DEP force. By utilizing the displacement difference, the separation of CTCs (modeled with A549 lung carcinoma cells) from RBCs inz‐axis direction was achieved. With the nonuniform electric field at optimized driving frequency, the RBCs were trapped in the cavities above the microchannel, whereas the A549 cells were separated with a high capture rate of 86.3% ± 0.2%. The device opens not only the possibility for 3D high‐throughput cell separation but also for future developments in 3D cell manipulation through rapid and low‐cost fabrication. 

   more » « less
4. Comparing Detection Schemes for Adversarial Images against Deep Learning Models for Cancer Imaging

   https://doi.org/10.3390/cancers15051548  

   Joel, Marina Z.; Avesta, Arman; Yang, Daniel X.; Zhou, Jian-Ge; Omuro, Antonio; Herbst, Roy S.; Krumholz, Harlan M.; Aneja, Sanjay (March 2023, Cancers)

   Deep learning (DL) models have demonstrated state-of-the-art performance in the classification of diagnostic imaging in oncology. However, DL models for medical images can be compromised by adversarial images, where pixel values of input images are manipulated to deceive the DL model. To address this limitation, our study investigates the detectability of adversarial images in oncology using multiple detection schemes. Experiments were conducted on thoracic computed tomography (CT) scans, mammography, and brain magnetic resonance imaging (MRI). For each dataset we trained a convolutional neural network to classify the presence or absence of malignancy. We trained five DL and machine learning (ML)-based detection models and tested their performance in detecting adversarial images. Adversarial images generated using projected gradient descent (PGD) with a perturbation size of 0.004 were detected by the ResNet detection model with an accuracy of 100% for CT, 100% for mammogram, and 90.0% for MRI. Overall, adversarial images were detected with high accuracy in settings where adversarial perturbation was above set thresholds. Adversarial detection should be considered alongside adversarial training as a defense technique to protect DL models for cancer imaging classification from the threat of adversarial images. 

   more » « less
5. Plasmonic-enhanced floating electrode optoelectronic tweezers (FEOET) for effective optical droplet manipulation

   S. Thio, S. Bae (January 2021, The 21st International Conference on Solid-State Sensors, Actuators and Microsystems (Transducers 2021))

   A plasmonic-enhanced floating electrode optoelectronic tweezers (FEOET) device is presented for effective optical droplet manipulation. Due to the importance of having a high-quality photoconductive layer, conventional FEOET devices face the issue between ineffective DEP performance and cost-ineffective fabrication. In this study, the use of metallic nanoparticles enables plasmonic light scattering to significantly enhance light absorption onto a photoconductive layer of the device, resulting in a largely improved dielectrophoretic (DEP) performance. Two numerical simulation studies have demonstrated the working principle of plasmonic-enhanced DEP and were further validated experimentally by an improved spectrophotometric light absorbance of the TiOPc layer, as well as demonstrating an 11-fold increase in light-actuated droplet speed. With much-improved DEP performance, this plasmonic-enhanced FEOET technology can provide a low-cost solution for various digital microfluidic (DMF) applications with the benefits of device simplicity. 

   more » « less