Kaoui, B. | Search bioRxiv

bioRxiv2025

Efficient calculation of fluid transport in porous media with moving boundaries

A novel hybrid model combining the lattice Boltzmann (LB) and finite difference (FD) methods is proposed to simulate transport in through a junction of actively contracting lymphatic vessels, while also handling flow of interstitial liquid in the surrounding porous tissue. Details of the dynamically flexing walls and valves in the lymphatic vessel and its near vicinity are modeled using a high-resolution LB method, whereas overall efficiency was significantly improved by using low-resolution FD in the larger tissue domain distant from the vessel. Pressure and velocity conditions at the interface between subdomains of the two numerical methods are matched by imposing a partial bounce-back ratio in LB corresponding to the permeability coefficient{kappa} in Darcys law for flow through porous media. Parameters governing the match between the algorithms at their interface can be estimated from the Kozeny-Carman relationship for porous media and further refined with a simpler, parallel flow geometry that also serves to validate the method. Test calculations show that the hybrid method is roughly four times faster than the LB method and permits computation over significantly larger domains. This method should be applicable to a large range of problems involving fluid flow in porous media with embedded conduits that have non-stationary boundaries. Author summaryIt is generally acknowledged that the finite difference method (FDM) is faster and requires less memory than the lattice Boltzmann method (LBM) for comparable domain sizes. However, LBM performs better for simulating fluid flow near complex, deformable, or moving boundaries. For this reason, it can be beneficial to create hybrid models that combine FDM and LBM. In this work, we use such a hybrid model to simulate a contracting lymphatic bifurcation in fluid. Our goal is to demonstrate the models robustness and high efficiency.

bioengineering↗

bioRxiv2025

Deep Learning Detection and Classification of Red Blood Cells: Towards a Universal Dataset

We evaluate emerging machine learning models for pattern recognition, focusing on the YOLOv11 architecture for detecting and classifying red blood cell shapes. Our analysis targets two characteristic morphologies observed under flow: slipper and parachute. A key challenge in this task is the development of a robust and diverse dataset. To address this, we employ synthetic image generation using a cut-and-paste approach, introducing variations in cell overlap and arrangements of microfluidic channels disposition to alleviate data scarcity and reduce cross-dataset bias. We generate these datasets with U-Net and Cellpose segmentation models, and rigorously assess YOLOv11 performance on two benchmarks: (i) a controlled dataset for evaluating classification accuracy, and (ii) a challenging, visually heterogeneous dataset for assessing generalization. Results show that the model achieves high precision for distinct cell types in controlled settings, but exhibits reduced performance on the unseen dataset, highlighting a trade-off between specialized accuracy and broad applicability in complex microscopy scenarios.

bioengineering↗