-
Notifications
You must be signed in to change notification settings - Fork 1
Web Tools
We have implemented several static javascript utilities to help visualize connectomic data in a flexible and light-weight manner. The first tool allows one to explore connectivity information in table form. The second tool, SharkViewer, is a 3D skeleton viewer. The third tool, Gorgonian, is a graph visualization tool. Finally, there is a simple web client for browsing the grayscale and segmentation data in DVID.
During the hackathon, these tools will be available through a Janelia server and will access the DVID server for data. In many cases, like for SharkViewer, the visualization does not require DVID but can work with specific data files, such as SWC files.
Stephan Gerhard imported the FIB-25 data and modified CATMAID to use DVID for grayscale tile serving. You can use his public server to inspect the data:
http://104.236.100.201/medulla/
The FlyEM Tables show statistics on the connectivity of cells by type.
FlyEM Tables are available during the hackathon.
FlyEM Tables is available on github under FlyEM Web Vizualizations. It can draw data from DVID using a key-value datastore or directly from JSON files.
SharkViewer is a web-based viewer for SWC neuron files. It is written entirely in JavaScript using Three.js. To view SWC format guidelines visit: http://research.mssm.edu/cnic/swc.html
An instance of SharkViewer is available with all of the skeletons from the 7 column reconstruction. You can view any neuron by selecting it's name or body id from the drop-down menu. Particle mode allows you to view a 3D representation quickly, while skeleton mode just shows the connections, not the nodes.
Outside of the hackathon you can set up your own instance of SharkViewer. Simply clone the SharkViewer github repository. To quickly view a swc file:
-
Open html/Shark_Viewer.html in a modern browser.
-
Click 'Choose File' button to load a SWC file.
More detailed instructions on how to customize your instance are available on the SharkViewer github repository.
Gorgonian is a web-based interactive graph visualization for the connectome. Gorgonian represents neuron as nodes and connections between neurons as edges. The user can select neurons and highlight primary and secondary connections. The minimum threshold for connection strength can be adjusted to view weaker or stronger connections interactively. The view plane can also be selected. Search functionality allowing a user to search for a neuron or group of neurons by id or name and allows regular expressions. Gorgonian is integrated with SharkViewer; at the click of a button you can view the highlighted neurons’ skeletons as well as the skeletons of the primary connections in 3D.
Gorgonian will be populated with the data from the seven column reconstruction during the hackathon.
Gorgonian is available on github under FlyEM Web Vizualizations. It can draw data from DVID using a key-value datastore or directly from JSON files.
You can watch this movie to learn how to use Gorgonian.
BTW: The name was inspired by the scientific name for sea fans.
The DVID Volume Viewer is a javascript, browser based, tool for viewing the grayscale imagery and segmentation data within a repository. It is similar in design to google maps, but with some additional features specific to this data type. For example, the imagery data can be interrogated in the XY, XZ, and YZ cut planes. There is also an interface to allow quick navigation to any point in the 3D volume. The intention is to facilitate quick investigation of the data without the need to install more complicated analysis tools.
A live instance of the Volume Viewer containing repositories from emrecon100 is available.
The volume viewer code is currently found as a stand alone webpage and can be checked out from github.
Table of Contents
- Overview
- Participants
- Workshop Schedule
- [Introduction to Fly Optic Lobe and Dataset](Introduction to Fly Optic Lobe and Dateset)
- Hackathon Goals
- Essential Graph Data
- Morphology Data
- DVID API
- Fly EM Analysis Tools
- Neutu
- [Web visualization](Web Tools)
- Hackathon Conclusions (links to relevant research)