Sequence of the Most Informative Joints (SMIJ): A new representation for human skeletal action recognition

By Ferda Ofli, Rizwan Ahmed Chaudhry, Gregorij Kurillo, René Vidal, Ruzena Bajcsy. 2012

404 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Artificial Intelligence

1. Jun Liu
2. G. Wang
3. Alex ChiChung Kot
4. Wanqing Li
5. Pichao Wang
6. Ling-yu Duan
7. Hossein Rahmani
8. Amir Shahroudy

In Biomedical Engineering

1. Jun Liu
2. G. Wang
3. Pichao Wang
4. Wanqing Li
5. Hossein Rahmani
6. Alex ChiChung Kot
7. Hong Liu
8. Qiuhong Ke

In Computer Vision and Pattern Recognition

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

Citations in the corpus

• Space-Time Representation of People Based on 3D Skeletal Data: A Review (2016, neutral): “[156] SMIJ Conc Lowlv Unsup X X X Miranda et al.”
• A novel hierarchical framework for human action recognition (2016, background): “A new representa-49 tion called SMIJ (sequence of the most informative joints) was proposed to select the most informative joints for performing an action [22], which was easy to interpret.”
• Tensor Representations via Kernel Linearization for Action Recognition from 3D Skeletons (2016, background): “Noteworthy are also parts-based approaches that additionally consider the connected body segments [30, 17, 16, 27].”
• 3D skeleton-based human action classification: A survey (2016, methodology): “Methods such as [64,63,97] focus on mining the subsets of most discriminative joints or consider the correlation of subsets of joints.”
• Deep Bilinear Learning for RGB-D Action Recognition (2018, background): “The temporal dynamics of each skeleton joint [15, 40, 5, 33] and joint pairs [41, 25, 20, 43, 29] are explored for mining the structure motions depicted in the skeleton sequences.”
• Bio-inspired Dynamic 3D Discriminative Skeletal Features for Human Action Recognition (2013, background): “[19] proposed the Sequence of Most Informative Joints (SMIJ) representation, a novel and highly interpretive feature for human motion representation for skeletal data based on joint angle time series.”
• 3D Flow Estimation for Human Action Recognition from Colored Point Clouds (2013, methodology): “Indeed, in (Ofli et al., 2012), a Sequence of Most Informative Joints (SMIJ) is computed based on measures like the mean of joint angles, the variance of joint angles and the maximum angular velocity of body joints.”
• Human Action Recognition From Various Data Modalities: A Review (2020, background): “[183] proposed to select the most informative representations based on the interpretable measures (e.”
• Jointly Learning Heterogeneous Features for RGB-D Activity Recognition (2015, methodology): “In the following, we define our notations first, and then present a detailed description of the proposed joint learning model.”
• Action recognition from motion capture data using Meta-Cognitive RBF Network classifier (2014, background): “Kurillo et al.[20] propose the use of histogram of most informative joints.”
• Real-time human action recognition from motion capture data (2013, background): “They form histogram of the number of occurrences of the k joints that are above certain threshold variance (most important joints).”