Corpus ID: 17728078

Feature sampling and partitioning for visual vocabulary generation on large action classification datasets

@article{Sapienza2014FeatureSA,
  title={Feature sampling and partitioning for visual vocabulary generation on large action classification datasets},
  author={M. Sapienza and Fabio Cuzzolin and P. Torr},
  journal={ArXiv},
  year={2014},
  volume={abs/1405.7545}
}
  • M. Sapienza, Fabio Cuzzolin, P. Torr
  • Published 2014
  • Computer Science
  • ArXiv
  • 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… CONTINUE READING

    Figures, Tables, and Topics from this paper.

    Deep Local Video Feature for Action Recognition
    • 59
    • PDF
    Feature sampling strategies for action recognition
    • 6
    • PDF
    Pooling the Convolutional Layers in Deep ConvNets for Video Action Recognition
    • 70
    • PDF
    Recognising and localising human actions
    • 4
    • PDF
    Beyond Gaussian Pyramid: Multi-skip Feature Stacking for action recognition
    • 226
    • PDF
    Long-short Term Motion Feature for Action Classification and Retrieval
    • 8
    • PDF

    References

    Publications referenced by this paper.
    SHOWING 1-10 OF 38 REFERENCES
    Action and Event Recognition with Fisher Vectors on a Compact Feature Set
    • 380
    • PDF
    Learning realistic human actions from movies
    • 3,374
    • Highly Influential
    • PDF
    Evaluation of Local Spatio-temporal Features for Action Recognition
    • 1,381
    • Highly Influential
    • PDF
    UCF101: A Dataset of 101 Human Actions Classes From Videos in The Wild
    • 2,540
    • Highly Influential
    • PDF
    Learning hierarchical invariant spatio-temporal features for action recognition with independent subspace analysis
    • 987
    • PDF
    HMDB: A large video database for human motion recognition
    • 1,834
    • Highly Influential
    • PDF
    Recognizing human actions: a local SVM approach
    • 2,829
    • PDF
    Learning discriminative space-time actions from weakly labelled videos
    • 49
    • PDF
    Towards optimal bag-of-features for object categorization and semantic video retrieval
    • 684
    • PDF
    Learning Discriminative Space–Time Action Parts from Weakly Labelled Videos
    • 63
    • PDF