DSPM: Dynamic Structure Preserving Map for action recognition

Cai, Qiao; Yin, Yafeng; Man, Hong

doi:10.1109/icme.2013.6607606

Cited by 21 publications

(14 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, even though a lot of approximate or progressive methods have been proposed to control the computational complexity, the application of graph matching methods is still limited. Cai et al [9,10] proposed to learn spatio-temporal dependency from low level features of similar images and reduce feature dimensions to improve the computation efficiency for both feature similarity and spatial smoothness. Some works formulate matching as an energy minimization problem.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Feature Guided Biased Gaussian Mixture Model for image matching

Sun

Tao

et al. 2015

Information Sciences

View full text Add to dashboard Cite

Section: Related Workmentioning

confidence: 99%

“…The EM algorithm is an elegant algorithm to solve the problem in Eq. (9). The EM algorithm alternates between two steps: the E-step and the M-step.…”

Section: The Solution Based On Em and Deterministic Annealingmentioning

confidence: 99%

Feature Guided Biased Gaussian Mixture Model for image matching

Sun

Tao

et al. 2015

Information Sciences

View full text Add to dashboard Cite

“…This method has become popular in recent years for data representation in various areas, such as bioinformatics (Zheng, Ng, Zhang, Shiu, & Wang, 2011) and computer vision (Cai et al, 2013). Recently, Cai, He, Han, and Huang (2011) argued that NMF fails to exploit the intrinsic local geometric structure of the data space.…”

Section: Introductionmentioning

confidence: 97%

Feature selection and multi-kernel learning for adaptive graph regularized nonnegative matrix factorization

Wang

Huang

Sun

et al. 2015

Expert Systems with Applications

View full text Add to dashboard Cite

a b s t r a c tNonnegative matrix factorization (NMF), a popular part-based representation technique, does not capture the intrinsic local geometric structure of the data space. Graph regularized NMF (GNMF) was recently proposed to avoid this limitation by regularizing NMF with a nearest neighbor graph constructed from the input data set. However, GNMF has two main bottlenecks. First, using the original feature space directly to construct the graph is not necessarily optimal because of the noisy and irrelevant features and nonlinear distributions of data samples. Second, one possible way to handle the nonlinear distribution of data samples is by kernel embedding. However, it is often difficult to choose the most suitable kernel. To solve these bottlenecks, we propose two novel graph-regularized NMF methods, AGNMF FS and AGNMF MK , by introducing feature selection and multiple-kernel learning to the graph regularized NMF, respectively. Instead of using a fixed graph as in GNMF, the two proposed methods learn the nearest neighbor graph that is adaptive to the selected features and learned multiple kernels, respectively. For each method, we propose a unified objective function to conduct feature selection/multi-kernel learning, NMF and adaptive graph regularization simultaneously. We further develop two iterative algorithms to solve the two optimization problems. Experimental results on two challenging pattern classification tasks demonstrate that the proposed methods significantly outperform state-of-the-art data representation methods.

show abstract

“…For the UCF dataset, our method achieved competitive accuracy compared to the statof-the-art. The best performance obtained by Cai et al [21] using dynamic structure preserving map (DSPM) technique. It, however, suffers from heavy computation and sensitivity to video data redundancy.…”

Section: Filter Training and Testingmentioning

confidence: 99%

“…Qualitative Results: Figure 3 ( Table 1 used SVM which is shown to be much slower than MCCF [12]. For memory usage, MCCF is very efficient, since the amount of memory required to learn Method Weizmann UCF sport Huang et al [19] 100% -Cai et al [21] 98.7% 90.6% Wang et al [17] 97.8 % 77.4% Campos et al [20] 96.7 % 80.0% Rodriguez et al [3] 86.6% 69.2% Yeffet & Wolf [22] -79.3% Our method 97.8% 82.6% an MCCF is independent of the number of training examples [12]. Whereas, the others suffer from memory overhead, as they need to load all training examples for learning.…”

Section: Filter Training and Testingmentioning

confidence: 99%

Multi-channel correlation filters for human action recognition

Kiani

Sim

Lucey

2014

2014 IEEE International Conference on Image Processing (ICIP)

View full text Add to dashboard Cite

In this work, we propose to employ multi-channel correlation filters for recognizing human actions (e.g. waking, riding) in videos. In our framework, each action sequence is represented as a multi-channel signal (frames) and the goal is to learn a multi-channel filter for each action class that produces a set of desired outputs when correlated with training examples. The experiments on the Weizmann and UCF sport datasets demonstrate superior computational cost (real-time), memory efficiency and very competitive performance of our approach compared to the state of the arts.

show abstract

DSPM: Dynamic Structure Preserving Map for action recognition

Cited by 21 publications

References 30 publications

Feature Guided Biased Gaussian Mixture Model for image matching

Feature Guided Biased Gaussian Mixture Model for image matching

Feature selection and multi-kernel learning for adaptive graph regularized nonnegative matrix factorization

Multi-channel correlation filters for human action recognition

Contact Info

Product

Resources

About