Refine your search:     
Report No.

Automatic similarity identification of 2D diffraction patterns with noisy background for 3D coherent X-ray diffractive imaging

Tokuhisa, Atsushi*; Jochi, Yasumasa*; Kono, Hidetoshi; Go, Nobuhiro*

We proposed a method for classifying and assembling two-dimensional diffraction patterns to reconstruct the three-dimensional diffraction intensity functions. In the classification process, similarity of each pair of patterns is judged based on the correlation of corresponding pixels along the concentric circles When two diffraction patterns are similar to each other, a straight line with high correlations appears from the origin to a certain high scattering-angle region. Theoretically, the integrated value of the correlation value along this line is proportional to the extent of the similarity and can be used as threshold for the classification. This correlation line disappears with increasing scattering-angle due to the quantum noise and the structural irregularity. In our estimation, diffraction images of 106orders will be required to construct the 3D structure with 3 ${AA}$; resolution for a molecule with a radius of 200 ${AA}$; In this case, the necessary number of pattern comparisons is $$sim$$ 10$$^{10}$$ order. To achieve such a huge number of comparisons, we have to develop an automatic system. In this meeting, we will report how to deal with such a huge amount of data and our current status of the system.



- Accesses





[CLARIVATE ANALYTICS], [WEB OF SCIENCE], [HIGHLY CITED PAPER & CUP LOGO] and [HOT PAPER & FIRE LOGO] are trademarks of Clarivate Analytics, and/or its affiliated company or companies, and used herein by permission and/or license.