Where PhDs and companies meet
Menu
Login

Already registered?

New user?

Example-based suggestion system for virtual film prototyping

ABG-96214 Master internship 6 months 546
2021-02-19
Logo de
LIX, Ecole Polytechnique
Palaiseau Ile-de-France France
  • Computer science
3D animation, virtual cinematography, deep learning
2021-03-05

Employer organisation

This work will take place at the GeoViC team (Geometry & Visual Computing) of LIX, at Ecole Polytechnique. If sanitary conditions allow, the work will take place on-site; otherwise remotely.

Description

Creating virtual films requires combining different shots taken from different cameras and viewpoints. This task is tedious and usually requires hours of manual watching clips and carefully selecting adequate combinations. To alleviate this, we are interested in developing techniques that enable fast editing of virtual films by exploiting real-life movies. Therefore, the primary goal of this thesis is to propose new deep-learning techniques that learn a mapping between an input filmed content and a directorial style (e.g. preferred viewpoints, motions, edits, or cutting rhythm). Given a new 3D animated content, the secondary goal is to learn how to generate a collection of interesting cinematographic suggestions. The successful candidate would need to work on deep-learning, virtual camera control and if needed filming conventions (with potential collaboration with the Pekin University and Beijing Film Academy).

Profile

Required skills

- In pursuit of Masters degree in a relevant field

- Experience computer graphics and in particular 3D animation

- Proven experience in Python

- High level of innovation and motivation

- Hands-on experience with deep learning frameworks (PyTorch)

- Communication skills in English

Starting date

2021-04-01
Partager via
Apply
Close

Vous avez déjà un compte ?

Nouvel utilisateur ?

More information?

Get ABG’s monthly newsletters including news, job offers, grants & fellowships and a selection of relevant events…

item1
They trusted us