Name
Abstract | ||
|---|---|---|
Cryo-electron microscopy (cryo-EM) technology has greatly facilitated the development of biology and medicine. Particle picking is a critical step in the processing of cryo-EM micrographs. However, achieving fast particle picking remains a bottleneck because the micrograph has a very low signal-to-noise ratio (below 0.1), large image size (usually 4k x 4k), small particle sizes and large numbers of particles. In this paper we propose a cGAN-based approach to mark out particle regions. We propose a data synthesis method to generate training samples thus there is no need to prepare particle samples from original micrographs. This data synthesis method will be very helpful when applying on different kinds of particle micrographs. We use the mean squared loss to improve the cGAN effect. In order to better demonstrate the performance of our method, we tested on the public dataset EMPIAR. The results show that our method can achieve fast and accurate automatic particle picking, and the performance is better than other known methods. |
Year | DOI | Venue |
|---|---|---|
2019 | 10.4310/cis.2019.v19.n3.a5 | COMMUNICATIONS IN INFORMATION AND SYSTEMS |
DocType | Volume | Issue |
Journal | 19 | 3 |
ISSN | Citations | PageRank |
1526-7555 | 0 | 0.34 |
References | Authors | |
0 | 5 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Fang Kong | 1 | 0 | 0.68 |
| Xirong Li | 2 | 1191 | 68.62 |
| Qing Liu | 3 | 31 | 2.95 |
| Chuangye Yan | 4 | 5 | 0.93 |
| Xinqi Gong | 5 | 6 | 2.80 |