I've been working on this problem on and off for over 10 years.
AMA ;-)
Are there any culling optimizations for unseen elements when layering SVG images? Looks like this isn’t an optimization that comes out-of-the-box with OpenVG and all the major web browsers needed to add this, so wondering what your solution is doing.
The depth buffer is used for opaque path interiors.