CIS 563: Physically based animation Ladislav Kavan Final Projects Final Project - What is it? A project of your choice - groups of 1-3
Download ReportTranscript CIS 563: Physically based animation Ladislav Kavan Final Projects Final Project - What is it? A project of your choice - groups of 1-3
CIS 563: Physically based animation Ladislav Kavan Final Projects Final Project - What is it? A project of your choice - groups of 1-3 people (of course we expect larger project from bigger team) can use existing code (but you have to clearly state it) you will also need to implement your own physically-basedrelated code discuss project ideas on Piazza – it’s a good idea to have the topic approved by instructors or TAs topics are not exclusive, multiple teams can work on the same topic 2 CIS563 Grade 55% homework 40% final project 5% class / piazza participation Bottom line: More important than grade is your demo reel Career: industry / academia (PhD) / startup 3 Important Dates 03/30: Teams Formed and Project Topic Decided. Submit to Canvas (everyone). 04/27: Final Projects due (writeup + code + video). Late days do not apply to final projects. 04/27 & 04/29: Final Project presentations (in class). 04/29: Last day of classes 4 Task 1: Project Topic & Team Project title & abstract (1 paragraph) Include names of all team members All team members are equal, all submit the same short document on Canvas 5 Previous Final Projects Ying: http://youtu.be/OC58YkYvmHQ Xinjie: https://youtu.be/DR5F7xzcCuo Sijie: https://www.youtube.com/watch?v=52_gbQDSmho 6 Projects can kickstart your thesis Research-type projects CIS 563 Final project can be used as springboard for Master’s thesis project Re-using your final project code, data, models etc. is allowed Unique opportunity to try research Without having to commit to a PhD program Nathan Marshak: http://youtu.be/cBVCbSq_7rI 7 Research-type projects Parallel Fast-mass Springs https://www.youtube.com/watch?v=N520KFOxaDg 9 Corotated elasticity using quaternions Might be more accurate than matrix-based corot elasticity due to the double cover property 10 Simulation of living tissues Surgical simulation - Very interesting materials - How to simulate e.g. connective tissues? https://www.youtube.com/watch?v=8tepFfxTaZw 11 iLarynx++: Intubation Simulator https://www.youtube.com/watch?v=d62JdxyscNI 12 Skinning Mesh Animations Approximate animation using skinning Used a lot in the industry Compression of pre-baked phys-based animations https://www.youtube.com/watch?v=e0rugcfR8K4 13 Production-type projects Topics: general recipe Pick a paper (typically SIGGRAPH or SCA) Variational integrators Cloth with self-collisions Rigid body simulation (will be covered in class) Finite element elasticity (will be covered in class) Control of articulated characters PIC and FLIP fluid simulation Level-set methods and water simulation Hair simulation Fire and explosions 15 Unified particle physics Based on PBD ideas, but more general https://www.youtube.com/watch?v=ffgRC3kvA_k 16 Symplectic integrators Better energy preservation Read related papers from Caltech (e.g. [Kharevych et al. 06]) Good for math/physics nerds 17 Cloth with self-collisions PBD or mass-spring Detect and respond to self-collisions Seminal paper: Robust Treatment of Collisions, Contact and Friction for Cloth Animation (Bridson, Fedkiw, Anderson, 2002) 18 Control of articulated characters Use existing rigid body simulator SIMBICON: Simple Biped Locomotion Control, Kangkang Yin, Kevin Loken, Michiel van de Panne SIGGRAPH 2007 More recent work by Stelian Coros 19 PIC and FLIP fluid simulation Hybrid particle/grid methods Used in industrial simulators Animating Sand as a Fluid, Yongning Zhu, Robert Bridson, SIGGRAPH 2005 20 Level set methods and water Practical Animation of Liquids, Nick Foster, Ronald Fedkiw, SIGGRAPH 2001 21 Hair simulation A Mass Spring Model for Hair Simulation, Selle, Lentine, Fedkiw, 2008 22 Fire and Explosions “Physically Based Modeling and Animation of Fire” by Nguyen, Fedkiw, Jensen 23 Make your own topic Look at recent SIGGRAPH papers http://kesen.realtimerendering.com/ Caveat: not always easy to implement Sometimes a lot of background is needed Ask for advice (how hard is it?) 24