This is the last class! Really great to meet everyone!! We covered anatomy-based animation, which takes super long time to render. It does not have a wide application in games and modern media for now but use the results to train a ml model sounds promising.

Crowd simulation Interesting concept: treat crowds as fluids. Not a good idea as the result shown in the video lol different models for simulation what would happen if two people are about to collide with each other.; Most research demonstrated are modelling people’s walk by assuming they walk to their destination in the shortest path, […]

The crowd simulation is very interesting. We could see the patterns of gathering and interaction in the crowd and it is pretty realistic. One thing I thought of is maybe we could add multiple state at the same time, say grouping while goal finding. It is quite similar to the multi hot embedding idea in […]

I feel like AI is the future of character animation. Before DNN we could hardly build smooth transitions for animations nor generate new movements from existing basis. Yet with methods like VAE and GAN, things become much easier. The trending methods such as diffusion may be also offering some inspiration in the field.

Everyone’s presentation is so interesting! I guess the main topic for most of the presentation today is to create stable and smooth transitions between animations. The one that use phase analysis opened my eyes

The character balancing algorithms are very interesting. The demo shown in class looks like the character really masters Karate lol. In the video of animals, we find many of the jumping instead of walking. I think in real life jumping is also a “very balanced” way of movement, but it costs more energy than walking, […]

I really like the bubble animations. I wish I could learn and understand more about MPM but it’s too hard lol

I’m really amazed by the presentation, which use AI to generate the sound in an animation sequence. Before that I never considered sound as part of animation pipeline. And there are actually very few technical artists working in the field. This paper opens up my mind and makes me really want to experiment on it […]

Not mush to say… It already takes some effort for me to understand the math behind the particle-based fluid simulation methods. I don’t have the ability yet to question or find any problems in it…. As for the presentation, it made me realize how important GPU parallelization is for cloth simulation. I could have a […]

I finally made decision to create AI based facial recognition. And Haoying would like to be my teammate.