Adaptive Tilt-Series Alignment With Feature Resampling in Cryo-Electron Tomography

Tilt-series alignment in cryo-electron tomography (cryo-ET) data processing is essential for visualizing high-quality structural information of macromolecules and organelles. Tilt-series alignment requires feature tracking, while the existing methods have challenges in identifying features in different kinds of specimens under a unified framework. In this paper, we propose an adaptive tilt-series alignment algorithm based on automatic feature seeking. With feature resampling based on local correlation coefficients, the proposed method can track distinguishable features in different kinds of specimens automatically. Experimental results show that the proposed method outperforms the existing methods on conventional specimens and yields comparable results on specimens with strong prior information from gold beads.