A Multiattribute Sparse Coding Approach for Action Recognition From a Single Unknown Viewpoint

Su, Te-Feng; Chiang, Chen-Kuo; Lai, Shang-Hong

doi:10.1109/tcsvt.2015.2409012

Cited by 14 publications

(11 citation statements)

References 70 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In popular low-dimensional manifold models [19], [53], for each feature space, one feature vector is represented by the linear combination of a few representative points. However, in these models the characteristics of the testing data can only be represented by the learned characteristics of one class from the training samples.…”

Section: A Preliminary: Cr-based Learning and Classificationmentioning

confidence: 99%

“…This results in larger within-class variations, making human action recognition more difficult than image classification tasks. Sparse representation (SR) has been successful for RGB-based action recognition [3], [19], [20]. The key assumptions of SR are: the features or representation of each category of training data is sufficient enough to span a separable subspace; and the training data are collected carefully, making the extracted feature space distribute uniformly.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Modal Human Action Recognition With Sub-Action Exploiting and Class-Privacy Preserved Collaborative Representation Learning

Liang

Liu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

Multimodal human action recognition with depth sensors has drawn wide attention, due to its potential applications such as health-care monitoring, smart buildings/home, intelligent transportation, and security surveillance. As one of the obstacles of robust action recognition, sub-actions sharing, especially among similar action categories, makes human action recognition more challenging. This paper proposes a segmental architecture to exploit the relations of sub-actions, jointly with heterogeneous information fusion and Class-privacy Preserved Collaborative Representation (CPPCR) for multi-modal human action recognition. Specifically, a segmental architecture is proposed based on the normalized action motion energy. It models long-range temporal structure over video sequences to better distinguish the similar actions bearing sub-action sharing phenomenon. The sub-action based depth motion and skeleton features are then extracted and fused. Moreover, by introducing within-class local consistency into Collaborative Representation (CR) coding, CPPCR is proposed to address the challenging sub-action sharing phenomenon, learning the high-level discriminative representation. Experiments on four datasets demonstrate the effectiveness of the proposed method.

show abstract

Section: A Preliminary: Cr-based Learning and Classificationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Multi-Modal Human Action Recognition With Sub-Action Exploiting and Class-Privacy Preserved Collaborative Representation Learning

Liang

Liu

et al. 2020

IEEE Access

View full text Add to dashboard Cite

show abstract

“…The current study is inspired by the investigation of techniques for multi-view action recognition using motion history images (MHIs) in combination with various descriptors and classifiers [34,[41][42][43]. In [41], multiple MHIs are computed using an overlapping temporal window, to avoid temporal discontinuities.…”

Section: Introductionmentioning

confidence: 99%

“…Major drawbacks of this approach are the large number of MHIs per video, the selection of optimal temporal window size and the training of a model without explicitly describing MHIs. The issue of the large number of MHIs per video has been addressed by constructing a multi interval MHI (MMHI) [42] using a fixed number of keyframes and temporal window size. The second drawback has been resolved by computing histogram of oriented gradients (HOG) [34,44] or histogram of directional derivative (HODD) [43] based description of single MHIs.…”

Section: Introductionmentioning

confidence: 99%

Multiple Batches of Motion History Images (MB-MHIs) for Multi-view Human Action Recognition

et al. 2020

View full text Add to dashboard Cite

The recognition of human actions recorded in a multi-camera environment faces the challenging issue of viewpoint variation. Multiview methods employ videos from different views to generate a compact view-invariant representation of human actions. This paper proposes a novel multi-view human action recognition approach that uses multiple low-dimensional temporal templates and a reconstruction-based encoding scheme. The proposed approach is based upon the extraction of multiple 2D Motion History Images (MHIs) of human action videos over nonoverlapping temporal windows, constructing Multiple Batches of Motion History Images (MB-MHIs). Then, two kinds of descriptions are computed for these MHIs batches based on (1) a deep Residual Network (ResNet) and (2) Histogram of Oriented Gradients (HOG) to effectively quantify a change in gradient. ResNet descriptions are average pooled at each batch. HOG descriptions are processed independently at each batch to learn a class-based dictionary using a K-Spectral Value Decomposition algorithm (K-SVD). Later, the sparse codes of feature descriptions are obtained using an Orthogonal Matching Pursuit (OMP) approach. These sparse codes are average pooled to extract encoded feature vectors. Then, encoded feature vectors at each batch are fused to form a final view-invariant feature representation. Finally, a linear Support Vector Machine (SVM) classifier is trained for action recognition. Experimental results are given on three versions of a multi-view dataset: MuHAVi-8, MuHAVi-14, and MuHAViuncut. The proposed approach shows promising results when tested for a novel camera. Results on deep features indicate that action representation by MB-MHIs is more view-invariant than single MHIs.

show abstract

“…After years of research, the face recognition technology in video has made great progress and development [5]. With the development of video surveillance, information security, access control and other application fields, videobased face recognition has become one of the most active research directions in the field of face recognition [6][7][8]. "Video-image" face recognition refers to the use of face video as input (query) to use the still image face database for recognition or verification [9].…”

Section: Introductionmentioning

confidence: 99%

Feature Extraction of Human Motion Video Based on Virtual Reality Technology

Zhou

2020

IEEE Access

View full text Add to dashboard Cite

In virtual reality scenes, the premise of moving target recognition in video is accurate target segmentation and extraction of target low-level features. In order to distinguish moving targets, it is not enough to use the underlying features. Further extraction of structural features that reflect the target structure can improve the recognition and tracking of moving targets. In order to ensure the stability of the feature areas extracted from the three-channel most stable extremum region, this paper proposes an improved algorithm of the three-channel most stable extremum region to improve the three-channel most stable extremum region. The algorithm can adaptively select the filters of each channel to filter the feature regions extracted from the three most stable extreme value regions. An action cycle is generally 30~50 frames, so it is faster and more advantageous to directly use the first 50 frames of video for processing. In this paper, two feature representation methods of variance gait energy graph algorithm and image splitting algorithm are proposed. The variance energy graph algorithm significantly improves the recognition rate; image splitting enhances the robustness of behavior classification. This paper proposes a feature representation algorithm of "distance from contour line to center line" to improve the recognition rate. By analyzing the feature extraction methods of principal component analysis, Fisher linear discriminant analysis and maximum divergence difference discriminant analysis, the main component discriminant analysis of row maximum divergence discriminant column and the two-dimensional two-dimensional maximum divergence discriminant analysis of row and column are proposed. This further enhances the ability to classify behaviors. INDEX TERMS human motion video; feature extraction; improved three-channel most stable extreme value area; virtual reality.

show abstract

A Multiattribute Sparse Coding Approach for Action Recognition From a Single Unknown Viewpoint

Cited by 14 publications

References 70 publications

Multi-Modal Human Action Recognition With Sub-Action Exploiting and Class-Privacy Preserved Collaborative Representation Learning

Multi-Modal Human Action Recognition With Sub-Action Exploiting and Class-Privacy Preserved Collaborative Representation Learning

Multiple Batches of Motion History Images (MB-MHIs) for Multi-view Human Action Recognition

Feature Extraction of Human Motion Video Based on Virtual Reality Technology

Contact Info

Product

Resources

About