Code for Particle Filter-based object tracking from color images. The framework utilizes importance sampling and dynamical system model for object tracking. It is capable of tracking in the presence of short-lived occlusion. The current implementation takes ground truth object location as input and infers object location when it is occluded.
- OpenCV 2.x (Included in repo)
For usage and demo please see main.cpp inside the project
TODO: Write Update, Predict, Resample function description and help for calling here
Further work can be done to write color histogram-based particles from the following paper:
- Nummiaro, Katja, Esther Koller-Meier, and Luc Van Gool. "An adaptive color-based particle filter." Image and vision computing 21.1 (2003): 99-110.
