Spatio-temporal Depth Cuboid Similarity Feature for Activity Recognition Using Depth Camera

By Lu Xia, Jake K. Aggarwal. 2013

459 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Representation & Self-Supervised Learning

How researchers see this paper

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

In Biomedical Engineering

1. Jun Liu
2. Wanqing Li
3. Pichao Wang
4. Hossein Rahmani
5. G. Wang
6. Mengyuan Liu
7. Amir Shahroudy
8. Qiuhong Ke

In Computer Vision and Pattern Recognition

1. Jun Liu
2. G. Wang
3. Luc Van Gool
4. Wanqing Li
5. Dacheng Tao
6. Ming-Hsuan Yang
7. Hossein Rahmani
8. Pichao Wang

In Representation & Self-Supervised Learning

1. Jun Liu
2. G. Wang
3. Amir Shahroudy
4. Yun Raymond Fu
5. Yu Kong
6. Pichao Wang
7. Wanqing Li
8. Hossein Rahmani

Citations in the corpus

• 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].”
• 3D Human Activity Recognition with Reconfigurable Convolutional Neural Networks (2014, methodology): “carefully designed features [36, 20, 35, 30].”
• Learning and Refining of Privileged Information-Based RNNs for Action Recognition from Depth Sequences (2017, neutral): “[51], Oreifej et al .”
• Unsupervised Learning of Long-Term Motion Dynamics for Videos (2017, methodology): “Typical features like HON4D [18], HOPC [19], and DCSF [40] capture spatio-temporal features in a temporal grid-like structure.”
• A Deep Structured Model with Radius–Margin Bound for 3D Human Activity Recognition (2015, methodology): “3% [49], 63 .”
• Are They Going to Cross? A Benchmark Dataset and Baseline for Pedestrian Crosswalk Behavior (2017, background): “These works either focus on recognizing the basic activity of a subject such as walking, running or jumping [41] or understanding the underlying semantics of a certain activity, e.”
• Improving Human Action Recognition Using Fusion of Depth Camera and Inertial Sensors (2015, methodology): “In [10], a filtering method extracted the spati otemporal interest points, followed by a depth cuboid simila rity feature for action recognition.”
• 3D-based Deep Convolutional Neural Network for action recognition with depth sequences (2016, background): “per normal vector (SNV) [2], actionlet [10], spatio-temporal depth cuboid [34], cloud points [35], etc.”
• Qualitative and Quantitative Spatio-temporal Relations in Daily Living Activity Recognition (2014, methodology): “Other approaches [22, 23] have used interest point detectors, extracted a 3D cuboid at each interest point, computed descriptors for each of the cuboids and then clustered similar descriptors together hence forming a feature descriptor codebook similar to the traditional bag-of-words approach.”
• ConvNets-Based Action Recognition from Depth Maps through Virtual Cameras and Pseudocoloring (2015, methodology): “The cross-subject scheme as the one in [11] was adopted which is different from the one adopted in [12] where more subjects were used for training in each round.”
• Human Action Recognition From Various Data Modalities: A Review (2020, methodology): “Therefore, many methods [193], [257] have attempted to fuse the skeleton and depth information for more accurate HAR.”
• Simplex-Based 3D Spatio-temporal Feature Description for Action Recognition (2014, unknown): “These desirable properties make local spatio-temporal features the most popular method to recognize actions, and continue to attract increasing attention from the computer vision community [7, 28].”