Differences
This shows you the differences between two versions of the page.
| Both sides previous revisionPrevious revisionNext revision | Previous revision |
| curriculum [2026/01/29 15:34] – respai-vic | curriculum [2026/07/02 15:08] (current) – [Period 1] respai-vic |
|---|
| |
| **CSC_53433_EP - Creative & Generative models in Computer Graphics (24h, 2 ECTS), Marie-Paule Cani (EP), Julien Pettré (Inria)** (contact: marie-paule.cani@polytechnique.edu) | **CSC_53433_EP - Creative & Generative models in Computer Graphics (24h, 2 ECTS), Marie-Paule Cani (EP), Julien Pettré (Inria)** (contact: marie-paule.cani@polytechnique.edu) |
| > This course presents the techniques for content creation in 3D Computer Graphics, from 3D shapes to animated landscapes and character motion, with a focus on their link with AI. We first introduce user-centered Creative AI (also called Expressive modeling methods), i.e. smart 3D models - either based on knowledge or trained from examples, designed to help users create/generate the 3D shapes their have in mind, and their organisation into complex, static or animated scenes. Second, we study the use of AI in Character Animation, from early motion planning and control methods to deep reinforcement learning solutions for generating the gaits of physically-based characters. These methods result into 3D agents able to navigate alone or within crowds, and to interact with their environment. The lab sessions are held on Unity, based on C#. | > This course reviews the recent advances of content creation methods in computer graphics — ranging from the design of 3D shapes to animated landscapes and character motion — with a focus on their connection to AI. We first present expressive modeling methods (also called "smart 3D models"), knowledge-based or learned from examples, designed to help users create the 3D shapes they have in mind, and arrange them into complex static or animated scenes. We then discuss their recent evolution toward user-centric creative AI, leveraging deep generative models. Next, we examine the generation and control of character motion, from navigation tasks with the presentation of recent crowd models, to new trends for motor skills generation, such as the use of Deep Reinforcement Learning for training motion controllers of physically-based characters. These methods enable the creation of 3D agents capable of moving individually or in groups while interacting with their environment. Practical sessions are conducted using Unity and the C# language. |
| | |
| |
