An Improved Correlation Method Based on Rotation Invariant Feature for Automatic Particle Selection
Particle selection from cryo-electron microscopy (cryo-EM) images is very important for high-resolution reconstruction of macromolecular structure. However, the accuracy of existing selection methods are normally restricted to noise and low contrast of cryo-EM images. In this paper, we presented an improved correlation method based on rotation invariant features for automatic, fast particle selection. We first selected a preliminary particle set applying rotation invariant features, then filtered the preliminary particle set using correlation to reduce the interference of high noise background and improve the precision of correlation method. We used Divide and Conquer technique and cascade strategy to improve the recognition ability of features and reduce processing time. Experimental results on the benchmark of cryo-EM images show that our method can improve the accuracy of particle selection significantly.
KeywordsParticle selection Rotation invariant feature Correlation Divide and Conquer Cascade strategy
Unable to display preview. Download preview PDF.
- 1.Joachim, F.: Three-dimensional electron microscopy of macromolecular assemblies. Academic Press (1996)Google Scholar
- 10.Abrishami, V., Zaldívar-Peraza, A., de la Rosa-Trevín, J., Vargas, J., Otón, J., Marabini, R., Shkolnisky, Y., Carazo, J., Sorzano, C.: A pattern matching approach to the automatic selection of particles from low-contrast electron micrographs. Bioinformatics 29(19), 2460–2468 (2013)CrossRefGoogle Scholar