Public--iconjobs.gifJob Offers[]

back to the job list

Jobs shown here are FOR HISTORIC INFO ONLY. The applicaiton deadline is EXPIRED

Offering Institution:Institut National de Recherche en Informatique et en Automatique, Sophia-Antipolis
Job Title:A Biologically Insprired Variational Approach for Multiple Motion Estimation: Variational Ap- proaches and Physcophysics
The visual perception of movement is important to virtually every living species. Neurophysiological
and psychophysical data [6] suggest that there several stages of motion analysis in the visual system. The motion response properties of neurons increase in complexity as one moves from primary
visual cortex (V1), up to higher cortical areas such as the middle temporal (MT) and the medial superior temporal area (MST). Many of the features of V1 neurons can now be replicated using computational models based on spatiotemporal filters. However until recently, relatively little was known about how the motion analysing properties of MT neurons could originate from the V1 neurons that provide their inputs.

In the Computer Vision community, there exists a very wide litterature on apparent motion estima-
tion, also called optical 
ow. Such a craze for optical 
ow is notably due by the number of applications that require some motion estimation to perform their tasks.We refer the reader to [8, 4, 1] for some reviews on this topic. Less models are proposed concerning multiple motions and it is our conviction that considering more complex stimuli will also bring some new solutions and ideas for simple optical 
ow estimation.

The candidate will work on two areas:

  - Variational approaches. Based on recent work in the team, the candidate will work on a variational approach (see [2] an overview) inspired by some of these biological facts, which integrates local velocities informations from V1 in order to compute wider patterns of motions as estimated in the MT area. We will focus on sequences presenting multiple local motions. The turbulent but directed 
ow of water in a stream or a 
ock of birds taking the air, are some examples. Motion tranparency is another special case since elements belonging to entities with different motions are interleaved or superimposed. From a perception point of view, it is noticeable that 150ms are needed to interpret and segment transparent sequences (compared to the 80ms for single motion), which suggest that more complex mecanisms are involved when more then one motion is present [7]. It is precisely this further processing performed by the visual cortex that we which to analyze and model via a variational formulation.

 - Psychophysics. An important part of the work will be to test the model on several visual stimuli used in psychophysics (see e.g. [3, 5] for transparent motion sequences). Interestingly, it is also expected that the analysis of recent perceptual studies will also enable us to improve the proposed variational formulation.

Technical requirements: PhD in computer science or allied field, preferably with a specialization on motion estimation
 - Mathematical and analytical abilities 
 - Good programming skills in C, C++

Interested candidates please contact: Pierre Kornprobst with cover letter, resume and a list of two potential references.

1. G. Aubert, R. Deriche, and P. Kornprobst. Computing optical 
ow via variational techniques. SIAM Journal
of Applied Mathematics, 60(1):156-182, 1999.
2. G. Aubert and P. Kornprobst. Mathematical Problems in Image Processing: Partial Differential Equations and
the Calculus of Variations, volume 147 of Applied Mathematical Sciences. Springer-Verlag, January 2002.
3. Oliver Braddick and Ning Qian. The organization of global motion and transparency. In Johannes M. Zanker
and Jochen Zeil, editors, Motion Vision - Computational, Neural, and Ecological Constraints. Springer Verlag,
Berlin Heidelberg New York,, 2001.
4. B. Galvin, B. McCane, K. Novins, D. Mason, and S. Mills. Recovering motion fields: an evaluation of eight
ow algorithms. British Machine Vision Conference, pages 195{204, 1998.
5. N. Grzywacz and E. Hildreth. The incremental rigidity scheme for recovering structure from motion: Position
vs. velocity based formulations. J. Opt. Soc. Amer. A., 4:503-518, 1987.
6. E.R. Kandel, J.H. Schwartz, and T.M. Jessel. Principles of Neural Science. McGraw-Hill, 4th edition, 2000.
7. G. Masson, D. Mestre, and L. Stone. Speed tuning of motion segmentation and discrimination. Vision Research,
39:4297-4308, 1999.
8. M. Orkisz and P. Clarysse. Estimation du 
flot optique en presence de discontinuitees: une revue. Traitement
du Signal, 13(5):489{513, 1996.
9. L.G. Ungerleider and M. Mishkin. Two cortical visual systems., pages 549-586. MIT Press, 1982.
(The job advertisement is expired. Therefore no contact info is shown)

NOTE: The list of job offers above is an informal synopsis for information purposes only and not legally binding. Please refer to the formal offers of the respective institution.


20 January 2010