Data for vidtrain

This training data consists of short video fluorescence microscopy sequences of microtubules crossing nanofabricated junctions. Annotations classify each junction crossing event at the moment the microtubule tip leaves the frame. Classes are: left->bottom, left->right, top->bottom, top->right Training data junction_stacks.npz: input stacks (shape: t, x, z, c) junction_annotations_2020-09-28_2036.npz: annotations (shape: t, c) junction_predicted_mean.npz: mean predictions from ensemble junction_predicted_std.npz: standard deviation predictions from ensemble Models network_* model weights, network parameters and models