Weighted averaging fusion for multi‐view skeletal data and its application in action recognition

Abstract

Existing studies in skeleton‐based action recognition mainly utilise skeletal data taken from a single camera. Since the quality of skeletal tracking of a single camera is noisy and unreliable, however, combining data from multiple cameras can improve the tracking quality and hence increase the recognition accuracy. In this study, the authors propose a method called weighted averaging fusion which merges skeletal data of two or more camera views. The method first evaluates the reliability of a set of corresponding joints based on their distances to the centroid, then computes the weighted average of selected joints, that is, each joint is weighted by the overall reliability of the camera reporting the joint. Such obtained, fused skeletal data are used as the input to the action recognition step. Experiments using various frame‐level features and testing schemes show that more than 10% improvement can be achieved in the action recognition accuracy using these fused skeletal data as compared with the single‐view case.

Distributed Human 3D Pose Estimation and Action Recognition

“…The third step is to perform the hybrid consensus iteration for L steps. For each iteration l, the i th node sends its prior informations (y l−1 i,k , Y l−1 i,k ), information contributions (u l−1 i,k , U l−1 i,k ) and b l−1 i,k to neighbor nodes and receives these consensus quantities from neighbors in parallel, then the consensus on this sensor node is performed according to (7), (8) and (9).…”

“…Current works mostly employ a centralized network topology, and use a central computer node to fuse information from all sensor nodes. The fusion algorithm can be a simple weighted summation method using skeleton joint tracking status [7], [8], [9], or additional bone length constrain conditions [10]. In order to use the temporal information and get smooth trajectory of skeleton joints, a Kalman filter or particle filter can be employed [11], [12], [13].…”

Multiple classifier-based spatiotemporal features for living activity prediction

“…Seo et al (2016) proposed an integrated technique to reject superfluous trajectories that resulted in camera motion, while still maintaining the effectiveness action prediction results. Some other groups focus on global feature descriptors for improving the result, such as Azis, Jeong, Choi, and Iraqi (2016) and Vishwakarma and Kapoor (2015). Azis et al (2016) proposed a method based on multi-view for action recognition using a weighted averaging fusion to merge skeletal data from multiple views.…”

“…Some other groups focus on global feature descriptors for improving the result, such as Azis, Jeong, Choi, and Iraqi (2016) and Vishwakarma and Kapoor (2015). Azis et al (2016) proposed a method based on multi-view for action recognition using a weighted averaging fusion to merge skeletal data from multiple views. Vishwakarma and Kapoor (2015) investigated a method hybrid classification model using SVM and k-Nearest Neighbour (k-NN) using human silhouette and grids for modelling feature vectors.…”

A Semi-blind Model with Parameter Identification for Building Temperature Estimation

“…In SAAs, the mean square deviation method [22,23] and weighted average method [24,25] are used. Therefore, the SAAs can be illustrated as the pseudo-code below.…”