view article

Figure 4
Using the output from the XChem CHiMP Detector Mask R-CNN model to determine coordinates to dispense compound for fragment soaking. For the mask images, low pixel values are depicted in blue and high pixel values are depicted in red. (a) Original micrograph of a crystallization drop. (b) The output bounding boxes, labels and masks from the XChem CHiMP Detector network. Bounding boxes are dark blue and associated object labels are in red. Instance-segmentation mask colours are randomized. (c) The predicted mask for the drop. (d) The mask for the drop minus the union of the predicted crystal masks. (e) The Euclidean distance transform of the mask in image (d); red pixels are those farthest from an edge. (f) The original image showing the coordinates determined by taking the maximum pixel value from image (e) as a red dot; this is the suggested dispensing coordinates.

Journal logoSTRUCTURAL
BIOLOGY
ISSN: 2059-7983
Follow Acta Cryst. D
Sign up for e-alerts
Follow Acta Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds