A post-doctoral position (18 months) is
available within the Video Processing and Perception Lab at Technicolor
Research & Innovation (Rennes, France).
Technicolor (http://www.technicolor.com) provides
technology, systems and services to its Media & Entertainment clients
involved in the different components of the video chain (content creation,
production, distribution and access). The Technicolor Research centers have developed
strong expertise in image and video processing and continuously invest in this
field. A main goal of our research is to anticipate incoming technological
evolutions, outperforming current limits. With this goal in mind, Technicolor
is investigating in the Human Perception and Video Processing areas. The
ambition is to understand how users perceive video (2D and 3D), derive
innovative computational models from this understanding and propose new
applications/services. Target applications include cinema, television, games,
web, video communication...
Technicolor Research & innovation in
Rennes, France, offers a Post-Doc position in the area of Human Perception and
Video Processing. The position is located in the Video Processing &
Perception Lab and in the project entitled “Human perception”. The
core technology consists in simulating the human visual attention, by detecting
the regions of interest (RoI) in images/videos. A biologically plausible
architecture based on signal processing tools has been implemented and is
currently restricted to bottom-up aspects of perception. The long-term ambition
is to extend the existing model towards top-down or more cognitive concepts of
attention with some tools, such as object modeling, specific detectors (face, text…)
or even scene categorization...
The context of this Post-doc is a
multi-modal visual attention model. More specifically, the main objective of
this research work is the use of audio saliency and/or audio cues in order to
extend and improve the existing Technicolor visual attention model [1,2].
Goals/tasks of this research work include
(other tasks to be discussed):
- A complete state-of-the-art in the field.
- A concrete implementation of a solution
within the existing model (C programming): development of the detection of
audio cues as well as their fusion with other independent visual cues.
- The definition/set-up of a complete test
environment in order to conduct experiments of visual attention with human
participants. It is foreseen that experiments will use an eye-tracking
apparatus in scenarios where visual stimuli are presented together with audio
stimuli.
The successful candidate must have a PhD
(or soon), and specific knowledge in Computer Science, Audio Processing and/or
Image Processing. Ideally, additional knowledge in Human Perception and a
background in subjective experiments setup/protocol would be valuable. Since
software development is expected in this research work, good programming skills
(C/C++ on Windows/Linux) are required.
Applicants should submit a curriculum
vitae, recent list of publications, a statement of research interests and
examples of research work achievements.
Resumes may be submitted electronically in
either Word (.doc), Rich Text (.rtf) or Portable Document Format (PDF), and
should be sent to Philippe.guillotel@xxxxxxxxxxxxxxx
[1] O. Le Meur, P. Le Callet, D. Barba, and
D. Thoreau, “A coherent computational approach to model the bottom-up
visual attention,” IEEE Trans. on Pattern Analysis and Machine
Intelligence, vol. 28, no. 5, pp. 802–817, 2006.
Dr. Quan HUYNH-THU
Senior Scientist, Video Processing & Perception Group
Technicolor Research & Innovation
email: quan.huynh-thu@xxxxxxxxxxxxxxx
tel: +33 (0)2 99 27 90 45
fax: +33 (0)2 99 27 30 15
|
|
Help preserve the color of our world
– Think before you print. |
