Thema
Multi-People Tracking Using Gait Cues
Typ Master, Forschungspraxis
Betreuer Maryam Babaee, M.Sc.
Tel.: +49 (0)89 289-28543
E-Mail: maryam.babaee@tum.de
Sachgebiet Computer Vision
Beschreibung In Multi-Object Tracking, the person recognition on realistic scenarios should be applied on videos with more than one person. This means that in addition to the identification task, the problem of tracking multiple objects has to be solved. In the case of multiple objects, entirely new problems arise such as mapping. This means that object observations must be assigned to the existing trajectories. For this assignment, the information like height, color and geometry distance are usually used to find a mapping between objects in consequent frames.

In this project, the Gait (the way of walking) is investigated as a biometric recognition feature. This is mainly suited for long-range surveillance such as on public squares, railway stations or airports. The early gait recognition systems are based on feature extraction followed by classification. As one of the most successful methods for basic feature extraction, we can mention the so-called Gait Energy Image (GEI) [1]. Here, the binary Silhouettes on a so-called Gait Cycle (i.e., a transition cycle of two steps) averaged.

In this work, the main focus would be on resolving one of the challenging problems in Multi Object Tracking (MOT) called ID-Switch where the assigned ID numbers of two persons crossing each other exchange along the tracking process. In other words, these two persons would be reidentified after a while. This undesirable occurrence could have negative effects in video surveillance applications. The goal is to use gait cues in order to track the people efficiently by reducing the number of ID-Switch.



Reference:
[1] J. Han und B. Bhanu, “Individual recognition using gait energy image”, IEEE Transactions on Pattern Analysis and Machine Intelligence, pp. 316-322, 2006.
[2] en.wikipedia.org/wiki/Gait_analysis
Voraussetzung It is expected that the prospective candidate has good Matlab and C++ programming skills as well as primary knowledge in graph optimization are required. For further questions and discussions, please contact me via email.
Bewerbung If you are interested in this topic, we welcome the applications via the email address above. Please set the email subject to “<Type of application> application for topic 'XYZ'”, ex. “Master’s thesis application for topic 'XYZ'”, while clearly specifying why are you interested in the topic in the text of the message. Also make sure to attach your most recent CV (if you have one) and grade report.