zSurgical

For my advanced surgical illustration course, I took on the challenge of constructing the sequence of images using Pixologic's ZBrush. I fell in love with this software very quickly and plunged myself into learning as many of it's features as possible to complete the main project for my surgical course. I was very fortunate to link up with an amazing surgeon at St. Michael's hospital in Toronto, who allowed me to observe and sketch during several spinal procedures. I have posted sketches from these OR experiences as well as the worked-up sequence sketches in previous posts.

I began building my 3D assets in Autodesk Maya, here you can see how I built the basic lumbar vertebrae using photo references. This allowed me to construct very low-polygon base meshes, which I then imported into ZBrush. (One drawback I've found in ZBrush is that it is not immediately obvious as to how to cut holes into a mesh). After increasing the subdivision levels, I could model fine details such as the cut portions of bone.

A unique modeling feature new to ZBrush is zSpheres, which I used to model the dura of the cauda equina and spinal nerves.

I wanted to re-create the entire surgical field in 3D assets, so that I could move my camera around to any vantage point and allow the audience to gain a better appreciation for the relative proportions of anatomical features.

I used a variety of "alphas" in zBrush to sculpt details on various tissue types. Alphas function very much like custom brushes in Photoshop, and can greatly enhance the organic quality of a 3D model.

As I put elements into place, I had to imagine how these structures would actually fit together. Certain anatomical features were difficult to find well-represented in reference images. In particular, I had a hard time finding a good illustration of the articular facet joint capsules.

For constructing the surgical tools, I returned to Maya and combined the modeling capabilities of both programs.

As I reached the final stages of fully modeled sequence shots, my machine really started to chug with all the high-resolution meshes. To compensate, I tried deleting lower sub-division levels so that I could keep the higher-levels with the finely sculpted details. Eventually I ended up exporting these to Maya and deleting faces to reduce the total poly count of the 3D scenes.

This ended up being a mistake in my process. What I should have done was convert these higher subdivision levels into displacement maps, but unfortunately I didn't figure out how to do this until later on down the road, even though the process turned out to be incredibly easy. Live and learn.

Figuring out the rendering system in ZBrush was another hurdle, but one I enjoyed, as it is a very satisfying stage to reach in the 3D workflow. I tried rendering in Maya as well, which allows for more camera control and lighting, but in the end I chose to stick with zBrush for the rendering. This project is still ongoing, but in the very last stage. It has already gone through several iterations in Photoshop and Illustrator. I will be submitting the final work to the AMI Salon at the upcoming conference in Baltimore, and I can't wait to get feedback from the plethora of talented artists in attendance.