Quantitative differential phase contrast imaging system using FPGA for cellular imaging

Abstract Quantitative differential phase contrast (qDPC) microscopy, a label-free method, allows cell analysis without needing labels. We introduce a compact, FPGA-based qDPC microscope module that enhances imaging efficiency by controlling essential functions such as digital illumination, rapid image capture, and phase reconstruction. This module integrates seamlessly with existing commercial microscopes, eliminating the need for additional optical components. A deep learning model is implemented to improve functionality, converting intensity images to phase data and requiring only a single image pair for isotropic phase reconstruction. Our approach holds promise for enhancing cell imaging and virology methodologies for epidemiological investigations. Our model accurately predicts quantitative phase maps, with a structural similarity index of ∼0.9 and a peak signal-to-noise ratio of ∼30. The FPGA-driven system achieves a tenfold increase in imaging speed, is advantageous for real-time cell imaging, and offers convenience through its unified programming language and hardware user interface. Additionally, our system shows significant potential for practical and scalable applications in basic research and clinical settings.