Improving Human Action Recognition Using Fusion of Depth Camera and Inertial Sensors

By Cheng Chen, Roozbeh Jafari, Nasser Kehtarnavaz. 2015

252 citations recorded by the source corpus.

Subfields: 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 Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

RGB-D datasets using microsoft kinect or similar sensors: a survey (2016, background): “The data modalities (the fifth column) include color, depth, skeleton and accelerometer, which are helpful for researchers to quickly identify the datasets especially when they work on multi-modal fusion [15, 16, 58].”
RGB-D-based action recognition datasets: A survey (2016, neutral): “Feature-Level-Fusion+SRC (Kinect+Acc1&Acc4) [96] 5.”
A survey of depth and inertial sensor fusion for human action recognition (2015, methodology): “In [13, 14, 43], it was shown that fusing information from depth and inertial sensors leads to more robust recognition.”
A Convolutional Neural Network-Based Sensor Fusion System for Monitoring Transition Movements in Healthcare Applications (2018, methodology): “Several action recognition sensor fusion systems have been developed by our research group in [8-12] by using two different modality sensors at the same time in order to achieve robust recognition under realistic operating conditions.”
Human Action Recognition From Various Data Modalities: A Review (2020, background): “[383] fused depth information and acceleration signals at both”
Fusion of depth, skeleton, and inertial data for human action recognition (2016, background): “A database consisting of 10 actions performed by 6 subjects is put together to carry out two types of testing of the developed fusion approach: subject-generic and subject-specific.”
3D skeleton-based human action classification: A survey (2016, methodology): “The method in [141] adopts a motion descriptor (similar to the motion history image in [142]) extracted from depth maps and acceleration statistics such as mean, variance, standard deviation and root mean square over a temporal window.”
3D action recognition using data visualization and convolutional neural networks (2017, background): “The development of RGB-D cameras, in particular the Kinect [5], opens up opportunities in addressing above problems [6, 7, 8, 9, 10].”
UTD-MHAD: A multimodal dataset for human action recognition utilizing a depth camera and a wearable inertial sensor (2015, methodology): “In [13], a fusion approach for improving human action recognition was developed based on depth and inertial sensors.”