Sparse-Scan Ptychography with Hybrid-Overlap Scanning and Reference-Guided Probe Updates
Sheng Chen, Zijian Xu, Ruoru Li, Xiangzhi Zhang, Renzhong TaiSparse-scan ptychography accelerates data acquisition but compromises the overlap-induced redundancy essential for stable blind reconstruction. In low-overlap cases, object errors are often transferred to the probe update, causing crosstalk and artifacts. In this work, we propose a sparse-scan ptychography method termed hybrid-scan ptychography, which combines a hybrid-overlap scanning strategy with a reference-guided ePIE reconstruction algorithm. In this method, a localized high-overlap scan is used to establish a reliable probe estimate for reconstructing a wide field of view scanned on a low-overlap grid. As a stabilizing prior, this probe estimate is integrated into the ePIE probe update for the low-overlap region through an additional quadratic regularization term, thereby ensuring accurate probe reconstruction for sparse-scan region. This method effectively suppresses sparse-scan artifacts and ensures robust, high-fidelity reconstruction, as demonstrated by numerical simulations and soft X-ray ptychography experiments. With the new method, the acquisition efficiency of ptychography can be increased by at least 2 times without significant resolution reduction.