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

This paper introduces a new generative deep learning network for human motionsynthesis and control. Our key idea is to combine recurrent neural networks(RNNs) and adversarial training for human motion modeling. We first describe anefficient method for training a RNNs model from prerecorded motion data. Weimplement recurrent neural networks with long short-term memory (LSTM) cellsbecause they are capable of handling nonlinear dynamics and long term temporaldependencies present in human motions. Next, we train a refiner network usingan adversarial loss, similar to Generative Adversarial Networks (GANs), suchthat the refined motion sequences are indistinguishable from real motioncapture data using a discriminative network. We embed contact information intothe generative deep learning model to further improve the performance of ourgenerative model. The resulting model is appealing to motion synthesis andcontrol because it is compact, contact-aware, and can generate an infinitenumber of naturally looking motions with infinite lengths. Our experiments showthat motions generated by our deep learning model are always highly realisticand comparable to high-quality motion capture data. We demonstrate the powerand effectiveness of our models by exploring a variety of applications, rangingfrom random motion synthesis, online/offline motion control, and motionfiltering. We show the superiority of our generative model by comparisonagainst baseline models.

Combining Recurrent Neural Networks and Adversarial Training for Human Motion Synthesis and Control