English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Human activity discovery aims to cluster the activities performed by humanswithout any prior information on what defines each activity. Most methodspresented in human activity recognition are supervised, where there are labeledinputs to train the system. In reality, it is difficult to label activitiesdata because of its huge volume and the variety of human activities. This paperproposes an unsupervised framework to perform human activity discovery in 3Dskeleton sequences. First, an approach for data pre-processing is presented. Inthis stage, important frames are selected based on kinetic energy. Next, thedisplacement of joints, statistical displacements, angles, and orientationfeatures are extracted to represent the activities information. Since not allextracted features have useful information, the dimension of features isreduced using PCA. Most methods proposed for human activity discovery are notfully unsupervised. They use pre-segmented videos before categorizingactivities. To deal with this, we have used a sliding time window to segmentthe time series of activities with some overlapping. Then, activities arediscovered by our proposed Hybrid Particle swarm optimization (PSO) withGaussian Mutation and K-means (HPGMK) algorithm to provide diverse solutions.PSO is used due to its straightforward idea and powerful global searchcapability which can identify the ideal solution in a few iterations. Finally,k-means is applied to the outcome centroids from each iteration of the PSO toovercome the slow convergence rate of PSO. The experiment results on fivedatasets show that the proposed framework has superior performance indiscovering activities compared to the other state-of-the-art methods and hasincreased accuracy of at least 4% on average.

A Novel Skeleton-Based Human Activity Discovery Using Particle Swarm Optimization with Gaussian Mutation