[visionlist] Postdoc position on Deep Learning and Computer Vision at KAIST, South Korea (editing of actors’ facial performance videos)

[jean-charles Bazin]

Topic: Postdoc position on Deep Learning and Computer Vision at KAIST,
South Korea, on editing of actors’ facial performance videos.



Job description: http://cml.kaist.ac.kr/hiring/CML_postdoc_pdf_face_v01.pdf

CML lab website: http://cml.kaist.ac.kr/

Prof. Jean-Charles Bazin: http://cml.kaist.ac.kr/pres_jcbazin.html

The Computational Media Lab (CML) led by Prof. Jean-Charles Bazin at KAIST
has one open postdoc position. The topic is Computer Vision and Deep
Learning for automatic editing of actor’s facial performance videos and
face reenactment. This research project will be conducted in collaboration
with Korea’s leading visual effects (VFX) studio.



#############################################

DETAILED JOB DESCRIPTION AND BENEFITS

==> http://cml.kaist.ac.kr/hiring/CML_postdoc_pdf_face_v01.pdf

#############################################



SALARY, TERMS AND BENEFITS
------------------------------------------

Salary: 4.5 millions KRW per month (around 4,000 US$).

Contract: one year (renewable, under conditions).

Starting date: September 2018 (flexible until December 2018).

Vacation: 20 days per year.



KAIST provides a wide range of benefits including sports facilities (gyms,
swimming pool, tennis courts, sauna, etc), accommodation, restaurants with
various choices, on-campus clinic, and many more.



Workplace: Computational Media Lab (CML) at KAIST, Daejeon, South Korea.
Daejeon is the fifth largest city of South Korea with a population of over
1.5 million. Daejeon is known as the Silicon Valley of Korea, and hosts
several public and private R&D centers, such as Samsung and LG.



Accommodation ranges from 200$ (on-campus dorm) to around 1000$ (large
family-apartment with 3 bedrooms outside the campus) per month.



Income tax is relatively low in Korea (especially compared to US and
Europe), and Korea has an excellent healthcare system (KAIST even has an
on-campus clinic).




JOB DESCRIPTION
-------------------------

1. You will conduct research on leveraging computer vision and deep
learning techniques for automatic editing of actor’s facial performance
videos and face reenactement. Possible research topics include:

-   Automatic synchronization of facial performance videos captured at the
same time (linear sync) and at different times (non-linear sync, see
[FaceDirector] and [ActionSnapping]).

-   Automatic video stabilization dedicated to facial performance videos.

-   Face reenactement in 2D and 3D (see [HeadOn] and [Face2Face]).

-   Investigation of deep learning for multi-modal data (audio+video).

-   Investigate visual editing (e.g. face manipulation) but also
sound/voice manipulation.



2. This research project will be conducted in collaboration with Korea’s
leading visual effects (VFX) studio, with state-of-the-art equipment and
cutting-edge technology.



3. You will be encouraged to publish in major venues of computer vision
(CVPR, ICCV, ECCV, TPAMI, IJCV), graphics (SIGGRAPH, TOG, TVCG) and machine
learning (ICML, NIPS).



4. You will have the opportunity to co-supervise interns, BS students, MS
students and PhD students.



*References*:

-   "FaceDirector: Continuous Control of Facial Performance in Video",
ICCV, 2015.

-   “HeadOn: Real-time Reenactment of Human Portrait Videos”, TOG, 2018.

-   “Face2Face: Real-time Face Capture and Reenactment of RGB Videos”,
CVPR, 2016.

-   "ActionSnapping: Motion-based Video Synchronization", ECCV, 2016



POSITION REQUIREMENTS

-----------------------------------
The candidate should have:

-    A PhD in computer science, engineering, mathematics, or other relevant
discipline (if thesis submission is pending, state expected submission
date).

-    A solid background in deep learning (especially CNN, RNN, LSTM and
GAN) and/or computer vision (especially 3D reconstruction from multiple
images and multi-view geometry).

-    Experience with OpenCV.

-    Experience in computer graphics is a plus (for example face
manipulation, blendshapes, and animation).

-    A solid publication record in top-tier venues of computer vision
(CVPR, ICCV, ECCV, TPAMI, IJCV), graphics (SIGGRAPH, TOG, TVCG,) and/or
machine learning (ICML, NIPS).





ABOUT KAIST

-------------------

KAIST is a world-leading public research university located in Daejeon,
South Korea. KAIST is recognized as a world-class research university. For
example, KAIST is ranked 14th in the world and 5th within Asia in
Engineering and Technology (QS World University Rankings). KAIST is also
ranked 6th in the world and 1st within Asia in the World's Most Innovative
Universities (Reuters). It is also ranked 5th of the world among the
universities that are aged 50 years or under (The Times Higher Education
Rankings).  KAIST is also a global university: more than 80% of major
courses are conducted in English, and international students from more than
70 different countries come to study at KAIST.



ABOUT CML

----------------

Website: http://cml.kaist.ac.kr/

The Computational Media Lab (CML) at KAIST is an interdisciplinary research
group dedicated to the analysis, processing and generation of media
contents with computer algorithms. Examples of our research topics include
video editing, AI (deep learning) for visual data, 3D telepresence,
augmented/virtual reality, and audiovisual media processing. We conduct
research in the fields of computer vision, computer graphics, robotics,
machine learning (deep learning) and multimedia. We invent novel algorithms
and create exciting new applications. We have collaborations and
connections with several universities (MIT, ETH Zurich, NTU Singapore) and
companies (Disney Research, Adobe Research, Samsung), as well as startups.
We publish papers in the premier venues of computer vision (TPAMI, CVPR,
ECCV and ICCV), graphics (TOG, SIGGRAPH, and SIGGRAPH Asia) as well as
robotics (IJRR).



The director of CML is Prof. Jean-Charles Bazin (
http://cml.kaist.ac.kr/pres_jcbazin.html). Before joining KAIST, he was an
Associate Research Scientist at Disney Research Zurich, and at the same
time an adjunct lecturer at ETH Zurich (2014-2016). Before joining Disney
Research, he was a Postdoc and a Senior Researcher at the Computer Graphics
Laboratory (Prof. Markus Gross) and at Computer Vision and Geometry Group
(Prof. Marc Pollefeys) at ETH Zurich (2011-2014). During this time, he was
a member of the Being There Center and was working in collaboration with
Disney Research Zurich. Before working at ETH, he was a Postdoctoral Fellow
at Computer Vision Lab of Prof. Katsushi Ikeuchi, Tokyo, Japan (2010/2011).
He received an MS in Computer Science at Universite de Technologie de
Compiegne, France (2006) and a PhD in Electrical Engineering at Robotics
and Computer Vision Laboratory, KAIST, South Korea (2011).





APPLICATION
------------------
If you want to apply or have any queries about the position, please contact
Prof. Jean-Charles (Director of CML) at bazinjc at kaist.ac.kr. Please provide:

-   CV with the list of publications.

-   A brief research statement (experience and interests).

-   Letter of recommendation (not mandatory, but appreciated).



The applications will start being reviewed on 25 August 2018 and will
continue until a successful candidate is selected.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://visionscience.com/pipermail/visionlist_visionscience.com/attachments/20180823/27771aab/attachment-0001.html>