These are examples of 8-frame 32x32 snippets that we produced from viHASi database. They are excerpted from longer 128x128 snippets of different computer-generated actors performing a "Kick" action. These are used in many of the simulations we are now running.  In fact, you can see the first one being processed by a Sparsey net on this site's homepage. And, see Experiment "32x32 Episodic Event Recognition" in "Results" menu. Most of these are clearly interpretable by humans, but at the same time, it might be hard to attribute any one particular high- or mid-level event category to some/most of the them. We are currently generating natural-based snippets like these, but transformed in certain well-defined ways and noisified so as to produce larger number of exemplar snippets where we have some control over intra- vs. inter-category variance. We are also doing the same thing for 64x64 snippets.