Title :
Combining Orientation Tensors for Human Action Recognition
Author :
Mota, Virginia F. ; Souza, Jessica I. C. ; de A Araujo, Arnaldo ; Bernardes Vieira, Marcelo
Author_Institution :
Dept. de Cienc. da Comput., Univ. Fed. de Minas Gerais, Belo Horizonte, Brazil
Abstract :
This paper presents a new tensor motion descriptor based on histogram of oriented gradients. We model the temporal evolution of gradient distribution with orientation tensors in equally sized blocks throughout the video sequence. Subsequently, these blocks are concatenated to create the final descriptor. Using a SVM classifier, even without any bag-of-feature based approach, our method achieves recognition rates greater than those found by other HOG techniques on KTH dataset and a competitive recognition rate for UCF11 and Hollywood2 datasets.
Keywords :
gradient methods; image motion analysis; image sequences; object recognition; video signal processing; HOG techniques; Hollywood2 datasets; KTH dataset; SVM classifier; UCF11 datasets; bag-of-feature based approach; competitive recognition rate; equally sized blocks; histogram of oriented gradients; human action recognition; orientation tensors; temporal gradient distribution evolution; tensor motion descriptor; video sequence; Computational modeling; Feature extraction; Histograms; Support vector machines; Tensile stress; Vectors; Video sequences; Histogram of gradients; Human action recognition; Motion description; Orientation tensor;
Conference_Titel :
Graphics, Patterns and Images (SIBGRAPI), 2013 26th SIBGRAPI - Conference on
Conference_Location :
Arequipa
DOI :
10.1109/SIBGRAPI.2013.52
