Fig. 1.

The overall scheme of TeraVR. a TeraVR is applicable to very challenging visualization and reconstruction scenarios such as complicated branching, weak signals, and overlapping neurites. With TeraVR, a user is able to combine stereoviews to observe the complex 3-D neurite patterns easily and perform the reconstruction effectively. Combining such visualization and data-exploration functions with terabyte-scale imaging data (e.g., whole-brain scale) management and streaming capability enables reconstruction of complex neuronal morphology at an optimized accuracy and efficiency. b A mixed reality visualization that demonstrates the use of TeraVR. Immersed in a virtual environment, the user manipulates the imaging data with TeraVR in a way similar to manipulating a physical object. c Multiple densely packed neurons (one in a different color) from an image with high, noisy background intensity level were reconstructed using TeraVR. d Real-time collaboration is demonstrated by showing views from all participating annotators. Each annotator logs onto the cloud and adopts a unique color for both annotation and an avatar representing the user’s real-time location. The left subpanel shows the view for annotator A (blue), in which two avatars of co-annotator B (purple) and C (yellow) are seen. Annotation results are instantly shared among them. The upper right subpanel: annotator B examined a partially traced segment by co-annotator C, only to identify more branches after turning up the contrast and having a close-up view of the segment (without affecting the views of other annotators); bottom right subpanel: the view of annotator C