Detecting Multiple Moving Objects in Crowded Environments with Coherent Motion Regions

Abstract

We propose an object detection system that uses the locationsof tracked low-level feature points as input, and producesa set of independent coherent motion regions as output. As an object moves, tracked feature points on it spana coherent 3D region in the space-time volume defined bythe video. In the case of multi-object motion, many possiblecoherent motion regions can be constructed around theset of all feature point tracks. Our approach is to identifyall possible coherent motion regions, and extract the subsetthat maximizes an overall likelihood function while assigningeach point track to at most one motion region. Wesolve the problem of finding the best set of coherent motionregions with a simple greedy algorithm, and show thatour approach produces semantically correct detections andcounts of similar objects moving through crowded scenes.

Cite

Text

Cheriyadat et al. "Detecting Multiple Moving Objects in Crowded Environments with Coherent Motion Regions." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2008. doi:10.1109/CVPRW.2008.4562983

Markdown

[Cheriyadat et al. "Detecting Multiple Moving Objects in Crowded Environments with Coherent Motion Regions." IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2008.](https://mlanthology.org/cvprw/2008/cheriyadat2008cvprw-detecting/) doi:10.1109/CVPRW.2008.4562983

BibTeX

@inproceedings{cheriyadat2008cvprw-detecting,
  title     = {{Detecting Multiple Moving Objects in Crowded Environments with Coherent Motion Regions}},
  author    = {Cheriyadat, Anil M. and Bhaduri, Budhendra L. and Radke, Richard J.},
  booktitle = {IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops},
  year      = {2008},
  pages     = {1-8},
  doi       = {10.1109/CVPRW.2008.4562983},
  url       = {https://mlanthology.org/cvprw/2008/cheriyadat2008cvprw-detecting/}
}