29
edits
Changes
Jump to navigation
Jump to search
- Input : Video data, 2D location While the tracking output of birds the Vicon system is in general very reliable, 3D trajectory (Labelled there are still mistakes and unlabeled)- Output: A. Create 2D missing data in the trajectories. Object detection on birdsThe idea is therefore to use visual tracking as an additional source of information to cope with e. (using simple blob tracking, or using machine learning)Bg. Match 3D - 2D trajectories markers which are missing due to identify identity flipsocclusion.C. Unlabeled trajectories can be combined with 2D trajectoriesSince the Vicon system allows to easily generate a large amount of ground truth data for the task of object detection, this is also a way to fill gaps train visual object detectors for the future and more "in 3D trajectories. (Project 2 output can be useful)-the-wild" scenarios.
Subprojects:
3.1 Offline data processing;
3.2 Online/quasi-real time solution working from the data stream and video stream (video stream is not directly available in real time for processing, but realtime view is generated in the software, so this could be grabbed)
Add name of and preferred method how to contact the main PI (i* Mate Nagy, mnagy@orn.empg. you)de* Hemal Naik, hnaik@orn.mpg.de* Bastian Goldluecke, bastian.goldluecke@uni-konstanz.de
List the aims of your The project, or what you expect anyone taking up essentially has two main parts. The first is to establish a pipeline for generating training data for visual object detectionusing the project Vicon system. The second part is supposed to hopefully achieveuse the trained detectors to augment the tracking, i.e. filling in gaps, helping with establishing identity, etc. The more specific Just like in other projects, both a high-quality offline solution as well as an online/quasi-real time solution working from the data stream and video stream would be desirable (the video stream is not directly available for processing, but a real-time view is generated in the bettersoftware, so this could be grabbed).
no edit summary
== Overview ==
== Contact ==
== Aims ==
Generating training data for visual object detection should be a pretty straight-forward standard problem, and is a good way to get into the project and data structures in the framework of a Bachelor/Master project.
An elaborate problem is to find a way to integrate the visual detections into the overall tracking pipeline, since this requires to find a suitable new algorithmic framework. It is closely related to [[Projects:Improve tracking of individual markers and marker patterns|this project]].Real-time is of course always harder.
== Provided data ==
The project uses [[Vicon:Data format documentation|data from the Vicon system]] to establish (partially labeled) 3D tracks, as well as input from RGB video cameras. Code for reading the data and calibration, as well as mapping 3D points to 2D images is available (TODO: put on CCU server one git server is up).
== Suggested/tested approaches ==
