Courses Offered

COMP3360 Data-driven computer animation

COMP3360 Data-driven computer animation

2022-23
Instructor(s):Komura Taku
(Class A) No. of credit(s):6
Recommended Learning Hours:
Lecture: 26.0
Tutorial: 13.0
Pre-requisite(s):COMP2119
Co-requisite(s):  
Mutually exclusive with:  
Remarks:

Course Learning Outcomes

1. [1]
understand the motivation and principles for character animation, as well as the skills for implementing such techniques
2. [2]
understand the motivation and principles of physical simulation
3. [3]
understand the motivation and techniques for learning character animation and physical simulation, implement solutions to learn from character motion data or physical animation data
Mapping from Course Learning Outcomes to Programme Learning Outcomes
 PLO aPLO bPLO cPLO dPLO ePLO fPLO gPLO hPLO iPLO j
CLO 1TTT
CLO 2TTT
CLO 3TTT

T - Teach, P - Practice
For BEng(CompSc) Programme Learning Outcomes, please refer to here.

Syllabus

Calendar Entry:
Basics of character animation, keyframe animation, motion capture, inverse kinematics, physically based character animation, Basics of physically-based animation, rigid body dynamics, point-based dynamics, hair animation, cloth simulation, facial animation, crowd simulation, mesh-shape editing, performance capture, skinning, data-driven character control, data-driven cloth animation, data-driven facial animation, data-driven skinning.

Detailed Description:

Introduction Mapped to CLOs
Course Introduction: Overview of character animation, overview of physical simulation, overview of shape deformation, Introduction to Python/NumPy. 1, 2, 3
Character Animation Mapped to CLOs
Character animation: Keyframe animation, motion capture, skinning, forward kinematics, inverse kinematics1
Facial animation: Blendshapes, musculoskeletal models, action units, face rigs, keyframe animation, video based capture 1
hand animation: video-based hand motion capture, hand rigging, collision detection, response 1
physically-based animation by pd control: Pd-control 1
crowd animation: Boids, reciprocal velocity obstacles, flow-based approaches 1
Data driven character animation Mapped to CLOs
Learning full-body character motion: Learning by LSTM, PFNN, local motion phase, RL for high level control 1, 3
Learning by track reference motion: reinforcement learning, deep reinforcement learning, learning physic-based animation 1, 3
learning facial animation: learning full face motion synthesis from sparse key point motion 1, 3
Learning crowd motion: Crowd animation synthesis by RL 1, 3
Physics based animation Mapped to CLOs
Point based dynamics: Cloth modeling, fluid modeling 2
Finite element model: Deformable models 2
Rigid body dynamics: Collision detection and response 2
Eulerian physics-based animation: MPM-based simulation, Eulerian fluid simulation 2
Learning Physics-based animation Mapped to CLOs
Learning physics by point-based methods2, 3
Learning shape deformation2, 3
Learning cloth deformation, Learning hair animation 2, 3

Assessment:
Continuous Assessment: 50%
Written Examination: 50%

Teaching Plan

Please refer to the corresponding Moodle course.

Moodle Course(s)

Please login with your CS account (for staff only)