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

Deep learning is ubiquitous across many areas areas of computer vision. Itoften requires large scale datasets for training before being fine-tuned onsmall-to-medium scale problems. Activity, or, in other words, actionrecognition, is one of many application areas of deep learning. While thereexist many Convolutional Neural Network architectures that work with the RGBand optical flow frames, training on the time sequences of 3D body skeletonjoints is often performed via recurrent networks such as LSTM.  In this paper, we propose a new representation which encodes sequences of 3Dbody skeleton joints in texture-like representations derived frommathematically rigorous kernel methods. Such a representation becomes the firstlayer in a standard CNN network e.g., ResNet-50, which is then used in thesupervised domain adaptation pipeline to transfer information from the sourceto target dataset. This lets us leverage the available Kinect-based data beyondtraining on a single dataset and outperform simple fine-tuning on any twodatasets combined in a naive manner. More specifically, in this paper weutilize the overlapping classes between datasets. We associate datapoints ofthe same class via so-called commonality, known from the supervised domainadaptation. We demonstrate state-of-the-art results on three publicly availablebenchmarks.

CNN-based Action Recognition and Supervised Domain Adaptation on 3D Body Skeletons via Kernel Feature Maps