Computational Algorithms for Automated Cell Density and Segmentation using Quantitative Phase Microscopy

Accurate cell confluence measurement is critical in biological experiments, yet traditional methods relying on visual inspection are subjective and inconsistent. We present a versatile Quantitative Phase Microscopy (QPM) system that automates confluence measurement and phase map analysis, providing a robust, non-invasive, and cost-effective solution. The system uses a common-path configuration with an azobenzene liquid crystal and a polarized camera to capture high-contrast intensity images. We evaluated the system on HeLa cell samples, achieving confluence measurement accuracies between 87%-98% for low, medium, and high-density samples. Automated cell segmentation was achieved using a computational algorithm that utilized block-matching and 3D filtering, k-means segmentation, cluster filtering, and morphological operations. This cost-effective tool operates with low-intensity light and resistance to vibrations, offering a robust, non-invasive solution for long-term cell monitoring, making it highly adaptable for research in both optical and biological fields.