Open Positions:

Post-Doctoral Position in Computer Vision, University of Amsterdam

CASSANDRA - Human Activity Recognition

The Intelligent Autonomous System (IAS) Group of the University of Amsterdam, The Netherlands, has an immediate opening for a Post-Doctoral Researcher in the area of computer vision and temporal data modelling.

The open position in the CASSANDRA project considers the topic of video-based human activity recognition in dynamic and cluttered environments. The particular application context is aggression detection in  surveillance, but generality to other application scenarios is desirable. Issues such as learning human appearance and motion, use of low-level features vs.  3D body pose recovery, tracking and movement classification are to be addressed. CASSANDRA furthermore investigates the fusion of video and audio information, and the role of context modelling. A unique realistic data set collected at a train station with professional actors and multiple synchronized cameras and microphones is available. See here for a description of work done on CASSANDRA so far.

We are looking for a highly motivated individual with an (existing or soon-expected) Ph.D. in the domain of computer vision and pattern recognition, preferably in the "Looking at People" area. The successful applicant will have a proven track record and be comfortable in formulating and pursuing research goals independently. The candidate will furthermore have good communication skills and able to collaborate effectively with other members of a research team. A certain affinity towards turning complex techniques into real-world practice (with the associated strong programming skills) is important. Good English skills are essential.

The open position involves an interesting research domain that provides rich publication potential. The Post-Doctoral Researcher will guide one or more PhD students and possibly several MS students, amplifying his/her impact. The University of Amsterdam hosts an internationally recognized vision group which offers ample opportunities for technical exchange.

Funding for this 2-year post-doctoral position is provided by the Dutch Science Foundation NWO. Monthly salary ranges from EUR 2708 to EUR 3442 (brutto) depending on experience. Funding prospects after 2 years are very good.

Inquiries are welcome.Prospective applicants please email

• cover letter (explaining suitability)
• resume
• copy of academic record
• supporting materials (e.g. PhD Thesis (draft), publications)

to Prof. Dr. D.M. Gavrila, contact info.
 

Ph.D. Position in Computer Vision, University of Amsterdam

Probabilistic Models for Human Appearance and Dynamics

The Intelligent Systems Lab Amsterdam (ISLA) of the University of Amsterdam, The Netherlands, has currently an opening for a Ph.D. student in the area of computer vision. The position concerns the challenging area of "Looking-at-People" which has important applications in public safety, elderly care and intelligent vehicles to human motion capture/analysis.

This Thesis will investigate probabilistic models for describing human appearance and dynamics. Work will concentrate on finding candidate body segments, and then on constructing assemblies of segments that are consistent  with the constraints on the appearance of a person that result from kinematic properties. The backdrop involves cluttered and dynamic environments, with possibly many people and sizable potential for occlusion.   

Admission for the Ph.D. position is competitive. Prospective applicants should demonstrate an above-average academic record and solid background in applied mathematics and computer science. Prior experience in computer vision and pattern recognition is a plus. The successful applicant will be able to conduct research independently, as well as collaborate effectively with other members of a research team. A certain affinity towards turning complex techniques into real-world practice is desired. Good English skills required.

The University of Amsterdam hosts an internationally recognized vision group. Funding for this Ph.D. student position is available for a 4-year period. Prospective applicants please email

• cover letter (explaining suitability)
• resume
• copy of academic record (courses/grades)
• supporting materials (e.g. MS Thesis, reports/publications, GRE/TOEFL scores)

to Prof. Dr. D.M. Gavrila, contact info

See also the Intelligent Systems Laboratory Amsterdam (ISLA) site.
 

Positions at the Master Level
Daimler Research or University of Amsterdam

Here are a few possible topics:

• Pedestrian Classification. We aim to detect humans in images by a pattern classification approach. Rather than trying to locate various body parts (e.g. head, hands, feet) in images explicitly based on prior knowledge about human appearance, we describe a region of interest in terms of low-level features and aggregate the latter into a feature vector. In an off-line training phase, a pattern classifier derives an internal pedestrian representation using a large numbers of previously categorized feature vectors. In the on-line recall phase the derived representation is used to classify unknown samples. Issues to be resolved are data normalization (e.g. size, contrast), feature selection (e.g. normalized pixel intensities, wavelets), dimensionality reduction (e.g. PCA, ICA) and actual pattern classification (e.g. SVM, NN).
  
  Extensive image data, both from the visible and infrared spectrum, is available from DaimlerChrysler. Existing MATLAB toolboxes for pattern classification can be used. System implementation for the recall phase is under C/C++, with emphasis on real-time considerations. The MS Thesis may lead to a publication.
   
• Deformable Contours (“Snakes”). Many of our techniques for object recognition involve classification techniques, which require large amounts of labeled data. The labeling task, in which a user for example outlines contours of certain objects in images, is not only tedious but also very time consuming. We want to develop computer vision techniques to assist the user in the above labeling task, allowing for some degree of automation. Of particular interest are so-called deformable contours which, after being placed over an image, adapt like "snakes" to better fit the image data. Objects of interest might be pedestrians, vehicles, faces, etc., but the resulting “PowerAssist” tool should be general-purpose.

Applicants should have an above-average academic record, a basic understanding of calculus and linear algebra, and possess strong programming skills (C/C++). Experience with image processing and/or pattern recognition techniques is a plus. Good English skills required, basic understanding of Dutch or German a plus.

Applicants should be available full-time for at least 5 months (longer is preferred). At DaimlerChrysler financial assistance is available. Prospective applicants please email

• cover letter (explaining suitability)
• resume
• copy of academic record (courses/grades)
• any supporting materials (project reports etc.)

to Prof. Dr. D.M. Gavrila, contact info.