Today we presented our Mini Project 2 projects. One of my favorite presentations was the one Sarah gave. She messed around with different skinning as well as different motion capture simulations. This was good to see since I wanted to do motion capture and skinning as well but left out of time. Plus I ran into some issues that Sarah did as well and she explained how she went around those problems. This will help me the next time I try using motion capture and skinning. But it also gave me ideas in terms of how to cred different kinds of crowd simulations as well.
One of my favorite papers that was presented today was the one that Sunil presented. The paper was called Spectral Style Transfer for Human Motion between Independent Actions. This paper focused on taking different kind of walking and jumping cycles and putting a certain style to it. For example if you put in jumping cycle than you can apply a childlike style to it to make it to act that way. It’s always interesting to compute style because they everyone has a different way of doing it, but it is still recognizable enough where anyone would recognize it.
One of the papers that we talked about was the SketchiMo: Sketch-based Motion Editing for Articulated Characters. I thought this was pretty cool because you could change the movement and flow of an animation by drawing a path and the avatar would follow that path. I thought overall they did good with not having too many IK or candy wrapper problems. I liked that instead of just having a proof of concept they also created a software to test. I want to follow up with this paper to see if I can try it out myself.
We talking more about skinning today. What I thought was interesting was that for skinning you have to also consider different kinds of weights to it, so it doesn't look too deformed. I liked the Pinocchio software that is used for skinning. The only downside to it is you have to use their models to use the outcome of different skinning models which I find to be pretty limiting.
We talked more about inverse kinematics, but this time we were shown different papers that tackle that. Again I always enjoy having a paper to reference to. Especially since I was familiar with authors such as Norm Badler, and Aaron Hertzmann. I didn't fully understand some of their papers in the past but I'm starting to recognize certain key terms and references from bring in this class. A paper that I was really interested in was the "Natural Motion Animation Through Constraining and Demonstrating At Will" I thought that it was interesting that one of the problems that IK tries to solve is a thing called the candy wrapper effect on limbs.
We talked about inverse kinematics today, This is a new topic to me. I didn't realize how inverse kinematics is a big component in making an animation be more fluid and realistic. The explanation was good, but I'm still trying to wrap my head around the math and thinking of different ways that I can implement this in unity to which it makes sense to me. I liked how people in robotics have to also think about this issue as well, which makes sense now that I think about it.