Group Sparsity and Geometry Constrained Dictionary Learning for Action Recognition from Depth Maps

By Jiajia Luo, Wei Wang, Hairong Qi. 2013

217 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Endocrinology, Diabetes and Metabolism

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Endocrinology, Diabetes and Metabolism

Citations in the corpus

NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis (2016, methodology): “The work of [22] utilized group sparsity based class-speciﬁc dictionary coding with geometric constraints to extract skeleton-based features.”
NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding (2019, methodology): “Some reported results on this dataset also achieved very high accuracies [40], [41], [42], [43].”
Deep Multimodal Feature Analysis for Action Recognition in RGB+D Videos (2016, methodology): “Recent single modality action recognition methods on depth signals can be divided into two major groups: depth map analysis methods [35, 41, 45, 79] and skeleton based methods [36, 65, 11, 73, 74].”
RGB-D-based action recognition datasets: A survey (2016, neutral): “alf training, half test) 2. CS (2,3,5,7,9 subject training, 1,4,6,8,10 subject test) 3. CS (1,3,5,7,9 subject training, 2,4,6,8,10 subject test) MSRDailyActivity3D [14] 1. τ-test [64] 2. DL-GSGC +TPM [65] 3. 3D joint+CS-MLtp [66] 4. Depth-VSFR [67] 1. 95.63 2. 95 3. 92.5 4. 89.7 1. D,S 2. S 3. C,S 4”
Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, methodology): “Bag-of-words encoding is widely employed by a large number of skeleton-based human representations [100], [106], [118], [134], [135], [138], [141], [143], [149], [152], [155], [157], [165], [170], [177], [179],”
Multimodal Multipart Learning for Action Recognition in Depth Videos (2015, methodology): “…of [5] Method (protocol of [5]) Accuracy Action Graph on Bag of 3D Points [5] 74.7% Histogram of 3D Joints [7] 79.0% EigenJoints [6] 83.3% Random Occupancy Patterns [8] 86.5% Depth HOG [38] 91.6% Lie Group [39] 92.5% JAS+HOG 2 [40] 94.8% DL-GSGC+TPM [41] 96.7% Proposed MMMP 98.2% into 88.13%.”
Skeleton-Based Human Action Recognition With Global Context-Aware Attention LSTM Networks (2017, background): “In the past few years, different feature extractors and classifier learning methods for skeleton based action recognition have been proposed [32]–[44].”
Skeleton-Based Action Recognition Using Spatio-Temporal LSTM Network with Trust Gates (2017, methodology): “A skeleton-based dictionary learning method using geometry constraint and group sparsity is also introduced in [12].”
Learning and Refining of Privileged Information-Based RNNs for Action Recognition from Depth Sequences (2017, background): “Action recognition from depth sequences [57, 34, 29, 44, 49] has attracted significant interest recently due to the emergence of low-cost depth sensors.”
Unsupervised Learning of Long-Term Motion Dynamics for Videos (2017, methodology): “State-of-the-art methods [15, 16] for RGB-D/depth-based activity recognition report human level performance on well-established datasets like MSRDailyActivity3D [14] and CAD-120 [33].”
Spatio-Temporal LSTM with Trust Gates for 3D Human Action Recognition (2016, methodology): “A skeleton-based dictionary learning utilizing group sparsity and geometry constraint was also proposed by [8].”
Jointly Learning Heterogeneous Features for RGB-D Activity Recognition (2015, background): “human skeleton (3D posture) tracker from single depth image [31], human motions can be effectively captured by the positional dynamics of each individual skeletal joint [7], [13], [23], [44] or the relationship of joint pairs [22], [27], [47] or even their combination [19], [52], [58].”
Bilinear heterogeneous information machine for RGB-D action recognition (2015, background): “Features from depth sequences can be encoded by [12], or be used to build actionlets [21] for recognition.”
Learning features combination for human action recognition from skeleton sequences (2017, methodology): “Many authors have proposed to extract spatial features from skeleton joints [23, 27], while others extract temporal information from sequences alignment [12] or by frequency analysis [24] of spatial features.”
Recognizing Actions in 3D Using Action-Snippets and Activated Simplices (2016, methodology): “…analysis (Xia, Chen, and Ag-garwal 2012), (ii) computing pairwise joint position features and using data mining to select the joints relevant to class labels (Wang et al. 2012b), and (iii) representing poses using a set of dictionaries which are learned by group sparsity (Luo, Wang, and Qi 2014).”
Linear-time online action detection from 3D skeletal data using bags of gesturelets (2015, methodology): “DL-GSGC adds group sparsity and geometric constraints to reconstruct feature points with minimal error.”
Real-Time Multi-scale Action Detection from 3D Skeleton Data (2015, methodology): “An example is the work by [17] where a sparse coding approach was applied to learn dictionaries for the action classes.”
Combined shape analysis of human poses and motion units for action segmentation and recognition (2015, unknown): “In [9], they differently use skeleton data by considering pairwise relative positions between joints.”
SkeletonNet: Mining Deep Part Features for 3-D Action Recognition (2017, methodology): “FTP has been combined with the SVM for action recognition [16], [18], [19].”