CO 417 - Advanced Computer Graphics:

Photographic Image Synthesis

Spring 2019

(Based on course offered at USC in Spring 2009, 2010)

 

Abhijeet Ghosh

Imperial College London

 



Description: W:\CO417_files\header.jpg

 

Administrativa:

Instructor

Abhijeet Ghosh    (ghosh@imperial.ac.uk)

TAs

 

 

Yuliya Gitlina (yuliya.gitlina13@imperial.ac.uk)

Yiming Lin (yiming.lin11@imperial.ac.uk)

Day and time

 

Monday 9:00am-11:00 am

Friday 2:00pm-4:00 pm

First class

January 14, 2019

Location

Hux 140 (Mon), 145 (Fri)

Office hours

Thursday 4:00-5:00pm, Hux 376

Prerequisites

CO 317 or equivalent

 

The final grade in this course will be based on two assignments (33% of grade) and a final exam (67% of grade).


 

Course Topics:

The course covers modern techniques for realistic image synthesis, with an emphasis on data-driven computer graphics. The objective of the course is to introduce the student to recent trends in advanced photographic image synthesis including image-based synthesis and alternative offline photorealistic rendering techniques, with applications in the entertainment industry.

Topics include high dynamic range imaging (HDRI), matting, bidirectional reflectance distribution functions (BRDFs), image-based relighting, global illumination, augmented reality, and computational photography. Furthermore, the course will also provide a brief introduction in the mathematical theory of Monte Carlo integration and the physics of light transport.

It is assumed that students taking the course already have had an introductory course in computer graphics or that they have done the equivalent reading. The following is a preliminary syllabus:

High Dynamic Range Imaging:

Image-based Lighting:

Light Fields:

         4D two-plane parameterization, acquisition, rendering.

Physics of light:

Sampling:

Global Illumination - Rendering:

Global Illumination - Advanced Material Models:

Image-based Material Representations:

Facial Capture and Reflectance:

Machine Learning for Rendering:

         neural networks applied to rendering and relighting applications.

 


 

Course Material:

The following textbooks are recommended but optional:

Lecture slides and links to additional reading material and papers will be made available.