Tracing using Omni

From Eyewire
Revision as of 03:34, 24 June 2016 by Pilnpat (Talk | contribs)

Jump to: navigation, search

What is Omni?

Omni is Eyewire's big brother. It is used by expert tracers in the lab. Omni uses larger data sets than Eyewire to render 3D images of neurons, but otherwise they are very similar, and drive towards the same goals.

There are a few differences between EyeWire and Omni. In Eyewire you see one 3D view and one 2D view which can be rotated to get a new orientation (alt+click+drag). In Omni all 3 2D views (along the X, Y, and Z axises) can be viewed simultaneously on multiple screens. Instead of a low opacity plane, Omni uses a navigational pointer to help orient the tracer in the 3D. There is also a second pointer that changes the pivot point in the 3D, so that the tracer can rotate the cube around whichever point is most convenient.

Mergers also pose a problem in Omni, just like they do in Eyewire. Omni has a little more flexibility when it comes to mergers however. Omni has a "break" function that allows the tracer to break the segments into smaller pieces. Many times breaking up segments created by the Artificial Intelligence (AI) corrects any problem merger. However, some segments still contain mergers even after being split, which must be traced around or left out of the final image.

Eyewire uses the consensus of multiple users to confirm the correct path of a given branch. Expert tracers are expected to have a high rate of accuracy. but every branch is still traced by two different users to account for human error.

We hope that by using Omni in the lab and Eyewire through our online community of amateur scientists, we can continue towards our goal of mapping the brain. Every day we're getting a little bit closer!

One OMNI cube is roughly the same "volume" as 125 EyeWire cubes.