Feature sampling and partitioning for visual vocabulary generation on large action classification datasets
By Michael Sapienza, Fabio Cuzzolin, Philip H. S. Torr. 2014
20 citations recorded by the source corpus.
The recent trend in action recognition is towards larger datasets, an increasing number of action classes and larger visual vocabularies. State-of-the-art human action classification in challenging video data is currently based on a bag-of-visual-words pipeline in which space-time features are aggregated globally to form a histogram. The strategies chosen to sample features and construct a visual vocabulary are critical to performance, in fact often dominating performance. In this work we provide a critical evaluation of various approaches to building a vocabulary and show that good practises do have a significant impact. By subsampling and partitioning features strategically, we are able to achieve state-of-the-art results on 5 major action recognition datasets using relatively small visual vocabularies.
Subfields: Computer Vision and Pattern Recognition
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Computer Vision and Pattern Recognition
- Alexander G. Hauptmann
- Zhenzhong Lan
- Bhiksha Raj
- Ming Lin
- Xuanchong Li
- Yi Yang
- Luc Van Gool
- Andrew Zisserman
Citations in the corpus
- Beyond Gaussian Pyramid: Multi-skip Feature Stacking for action recognition (2014, background): “[28] explored ways to subsample and generate vocabularies for Dense Trajectory features.”