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

Drone-camera based human activity recognition (HAR) has received significantattention from the computer vision research community in the past few years. Arobust and efficient HAR system has a pivotal role in fields like videosurveillance, crowd behavior analysis, sports analysis, and human-computerinteraction. What makes it challenging are the complex poses, understandingdifferent viewpoints, and the environmental scenarios where the action istaking place. To address such complexities, in this paper, we propose a novelSparse Weighted Temporal Fusion (SWTF) module to utilize sparsely sampled videoframes for obtaining global weighted temporal fusion outcome. The proposed SWTFis divided into two components. First, a temporal segment network that sparselysamples a given set of frames. Second, weighted temporal fusion, thatincorporates a fusion of feature maps derived from optical flow, with raw RGBimages. This is followed by base-network, which comprises a convolutionalneural network module along with fully connected layers that provide us withactivity recognition. The SWTF network can be used as a plug-in module to theexisting deep CNN architectures, for optimizing them to learn temporalinformation by eliminating the need for a separate temporal stream. It has beenevaluated on three publicly available benchmark datasets, namely Okutama,MOD20, and Drone-Action. The proposed model has received an accuracy of 72.76%,92.56%, and 78.86% on the respective datasets thereby surpassing the previousstate-of-the-art performances by a significant margin.

SWTF: Sparse Weighted Temporal Fusion for Drone-Based Activity Recognition