As somebody who has been doing such tasks in the 80ties-90ties (reconstructing color specific connectivity of primate cone photoreceptors to so-called horizontal cells and bipolar cells by making transparent sheet overlays from hardcopy photo-montages of TEM serial sections) I can appreciate the significant step of your approach.
It took us months to get a few cells done. So what has come more or less to an halt due to the almost prohibitive workload involved - HiRes reconstruction of synaptic architecture and connectivity - gets a new chance by opening this joint effort (although I share the concerns of deanhorak regarding the limits of identifying the finer synaptic structures).
So, my obvious question is: Does your volume of sections include the outer retinal layers and their cell types, and/or do you have any intention to include them? (I can understand, if you consider focusing on the directionally sensitive elements in the inner retina is more than enough)
No is the answer, unfortunately. The dataset we are dealing with on EyeWire contains only the inner plexiform layer where all the interesting things (for us, of course) happen and the two nuclear layers whose cells send their neurites to it.
On the other hand, we have another dataset, whose patch area is a little bit smaller than that of EyeWire’s, but contains all the layers including the photoreceptors. We don’t have any specific plan at the moment regarding what to do with it.