Vision-guided parallel manipulation of cells with optoelectronic tweezers

Cell manipulation in microfluidic systems is essential for applications such as single-cell analysis, tissue engineering, and drug screening. Optoelectronic tweezers enable flexible, scalable, and non-contact cell manipulation. However, achieving automated closed-loop manipulation of multiple cells while avoiding collisions remains challenging in real microfluidic environments. Here, we present VPM-OET (Vision-guided Parallel Manipulation Optoelectronic Tweezers), a vision-guided closed-loop framework for robust parallel cell manipulation. The system combines computer vision-based detection, dynamic multi-target tracking, and collision-free trajectory planning to continuously monitor cell motion and update manipulation paths during operation. Using yeast cells and mammalian cells in microfluidic chips as model systems, we demonstrate cell transport around an obstacle region, parallel manipulation of 10 cells, and collision-free cross-path manipulation in separate experiments. This work offers a practical strategy for automated closed-loop OET control and expands the potential of parallel cell handling in microfluidic and biomedical applications.