A VIDEO TENSOR SELF-DESCRIPTOR BASED ON VARIABLE SIZE BLOCK MATCHING
Keywords:
self-descriptor, compact descriptor, variable size block matching, human ac- tion recognitionAbstract
This paper presents a dierent and simple approach for video description using only block matching vectors, considering that most works on the eld are based on the gradient of image intensities. We rst divide the image into blocks of dierent sizes. The block matching method returns a displacement vector for each block, which we use as motion information to obtain orientation tensors and to generate the nal self-descriptor, since it depends only on the input video. The resulting descriptor is evaluated by a classication of KTH, UCF11 and Hollywood2 video datasets with a non-linear SVM classier. Our results indicate that the method runs fast and has fairly competitive results compared to similar approaches. It is suitable when the time response is a major application issue. It also generates compact descriptors which are desirable to describe large datasets.
Downloads
References
Figueiredo, Ana MO and Maia, Helena A and Oliveira, Fabio LM and Mota, Virgnia F and
Vieira, Marcelo Bernardes (2014) A Video Tensor Self-descriptor Based on Block Matching, Computational
Science and Its Applications{ICCSA 2014, pp 401{414.
Ha ane, Adel and Palaniappan, Kannappan and Seetharaman, Guna (2008), UAV-Video Registra-
tion Using Block-based Features, IEEE International Geoscience and Remote Sensing Symposium
(IGARSS), Vol. 2, pp. 1104-1107.
Amel, Abdelati Malek and Abdessalem, Ben Abdelali and Abdellatif, Mtibaa(2010), Video Shot
Boundary Detection Using Motion Activity Descriptor, Journal of Telecommunications, Vol.2, pp.
-59.
Sun, Xinding and Divakaran, Ajay and Manjunath, BS (2001), A Motion Activity Descriptor and
its Extraction in Compressed Domain, Proceedings of the 2nd IEEE Paci c Rim Conference on
Multimedia, Springer, pp 450-457.
Chan, MH and Yu, YB and Constantinides, AG (1990), Variable size block matching motion
compensation with applications to video coding, IEE Proceedings I (Communications, Speech and
Vision), IET, Vol.137, pp 205-2012.
Horowitz, Steven L and Pavlidis, Theodosios (1976),Picture segmentation by a tree traversal al-
gorithm, Journal of the ACM (JACM), ACM, Vol.23, pp 368-388.
Mota, Virginia F and Souza, Jessica IC and Araujo, Arnaldo de A and Vieira, Marcelo Bernardes
(2013), Combining Orientation Tensors for Human Action Recognition, Conference on Graphics,
Patterns and Images (SIBGRAPI), pp 328-333.
Sad, Dhiego and Mota, Virginia Fernandes and Maciel, Luiz Maurlio and Vieira, Marcelo
Bernardes and Araujo, Arnaldo de A (2013), A Tensor Motion Descriptor Based on Multiple
Gradient Estimators, Conference on Graphics, Patterns and Images (SIBGRAPI),pp 70-74.
Perez, Eder de Almeida and Mota, Virgnia Fernandes and Maciel, Luiz Maurlio and Sad, Dhiego
and Vieira, Marcelo Bernardes (2012) , Combining Gradient Histograms Using Orientation Tensors
for Human Action Recognition, 21st International Conference on Pattern Recognition (ICPR), pp
-3463.
Puri, A and Hang, HM and Schilling, DL (1987), Interframe coding with variable block-size motion
compensation, Globecom, Vol. 87, pp 65-69.
Muralidhar, P and Rao, CB Rama and Kumar, I Ranjith (2012) , Ecient architecture for variable
block size motion estimation of h. 264 video encoder, International Conference on Solid-State and
Integrated Circuit (ICSIC), Vol.32.
Kim, JongWon and Lee, Sang Uk (1994),Hierarchical variable block size motion estimation tech-
nique for motion sequence coding, Visual Communications' 93, International Society for Optics
and Photonics, pp 372-383.
Rhee, Injong and Martin, Graham R and Muthukrishnan, S and Packwood, Roger A (2000),
Quadtree-structured variable-size block-matching motion estimation with minimal error, Circuits
and Systems for Video Technology, IEEE Transactions on, Vol.10, pp 42-50.
Ji, Yanli and Shimada, Atsushi and Taniguchi, R-i (2010), A Compact 3D Descriptor in ROI for
Human Action Recognition, IEEE TENCON, pp 454-459.
Mota, Virgnia Fernandes and Perez, Eder de Almeida and Vieira, Marcelo Bernardes and Maciel,
LM and Precioso, Frederic and Gosselin, Philippe Henri (2012), A Tensor Based on Optical
Flow for Global Description of Motion in Videos, Conference on Graphics, Patterns and Images
(SIBGRAPI), pp 298-301.
Alexander Klaser and Marcin Marsza lek and Cordelia Schmid (2008), A Spatio-Temporal Descrip-
tor Based on 3D-Gradients, British Machine Vision Conference (BMVC), September, pp 995-1004.
Mota, Virginia Fernandes and Perez, Eder de Almeida and Maciel, Luiz Maurlio and Vieira,
Marcelo Bernardes and Gosselin, Philippe-Henri (2013), A Tensor Motion Descriptor Based on
Histograms of Gradients and Optical Flow, Pattern Recognition Letters, Vol.31, pp 85-91.
Po, Lai-Man and Ma, Wing-Chung (1996), A Novel Four-step Search Algorithm for Fast Block
Motion Estimation, IEEE Transactions on Circuits and Systems for Video Technology, Vol. 3, pp
-317.
Nie, Yao and Ma, Kai-Kuang (2002), Adaptive Rood Pattern Search for Fast Block-matching Mo-
tion Estimation, IEEE Transactions on Image Processing, Vol.11, pp1442-1449.
Zhu, Shan and Ma, Kai-Kuang (2000), A New Diamond Search Algorithm for Fast Block-matching
Motion Estimation, IEEE Transactions on Image Processing, Vol.9, pp 287-290.
Johansson, Bjorn and Farneback, Gunnar (2002), A theoretical comparison of di erent orientation
tensors, Proceedings SSAB02 Symposium on Image Analysis, Citeseer, pp 69-73.
Schuldt, Christian and Laptev, Ivan and Caputo, Barbara (2004), Recognizing Human Actions: a
Local SVM Approach, Proceedings of the 17th International Conference on Pattern Recognition
(ICPR), Vol.3, pp 32-36
Liu, Jingen and Luo, Jiebo and Shah, Mubarak (2009), Recognizing Realistic Actions from Videos
in the wild, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), pp 1996-2003
Marszalek, Marcin and Laptev, Ivan and Schmid, Cordelia (2009), Actions in Context, IEEE
Conference on Computer Vision and Pattern Recognition (CVPR), pp 2929-2926.
Torralba, Antonio and Fergus, Robert and Weiss, Yair (2008) Small Codes and Large Im-
age Databases for Recognition, IEEE Conference on Computer Vision and Pattern Recognition
(CVPR).
Wang, Heng and Klaser, Alexander and Schmid, Cordelia and Liu, Cheng-Lin (2013), Dense
Trajectories and Motion Boundary Descriptors for Action Recognition, International Journal of
Computer Vision, Vol.103, pp 60-79.
Wang, Heng and Schmid, Cordelia and others (2013), Action recognition with improved trajectories,
International Conference on Computer Vision.
