Synthetic patches, real images: screening for centrosome aberrations in EM images of human cancer cells

08/27/2019
by   Artem Lukoyanov, et al.
17

Recent advances in high-throughput electron microscopy imaging enable detailed study of centrosome aberrations in cancer cells. While the image acquisition in such pipelines is automated, manual detection of centrioles is still necessary to select cells for re-imaging at higher magnification. In this contribution we propose an algorithm which performs this step automatically and with high accuracy. From the image labels produced by human experts and a 3D model of a centriole we construct an additional training set with patch-level labels. A two-level DenseNet is trained on the hybrid training data with synthetic patches and real images, achieving much better results on real patient data than training only at the image-level. The code can be found at https://github.com/kreshuklab/centriole_detection.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset