Quantitative differential phase contrast phase reconstruction for sparse samples

Abstract

Quantitative differential phase contrast (qDPC) microscopy can achieve phase imaging of unlabeled cell samples with high spatial resolution. However, qDPC imaging is easily affected by the noise generated in the experiments. In this manuscript, the L0-norm regularization was introduced into qDPC phase reconstruction for sparse samples (L0-qDPC), including the quantitative phase target and cell samples. After the detailed comparison of phase fidelity, imaging contrast, resolution, and convergence rate, L0-qDPC method can provide a stable qDPC phase imaging without parameter adjustment for sparse samples due to strong constraint and good robustness of L0-norm based on sparse prior, compared with the L2-norm and total variation regularization.