Projects
-
Person Re-Identification
For many applications such as person tracking or person retrieval it is not necessary to uniquely identify a person, i.e. assign a unique name or identity. It often suffices to determine previous or future occurrences of the same person in other images or videos. Person re-identification deals with the problem of recognizing people given a textual description, a sample image or video of the person.
-
Face and Person Tracking
In unconstrained scenarios a necessary pre-condition to person (re-)identification is the localization of faces and persons. Tracking allows to both localize faces and persons and associate them over time. We target mostly unconstrained environments, i.e. neither assume static backgrounds nor restrict to frontal poses.
-
A Universal Labeling Tool for Computer Vision: Sloth
Sloth's purpose is to provide a versatile tool for various labeling tasks in the context of computer vision research. Since there are so many different label formats and requirements out there, we concluded that is virtually impossible to build the one label tool sufficient to handle all labeling tasks. Therefore, this project can be seen rather as a framework and set of standard components to quickly configure a label tool specificly tailored to ones needs.