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

With the fast development of AI-related techniques, the applications oftrajectory prediction are no longer limited to easier scenes and trajectories.More and more heterogeneous trajectories with different representation forms,such as 2D or 3D coordinates, 2D or 3D bounding boxes, and evenhigh-dimensional human skeletons, need to be analyzed and forecasted. Amongthese heterogeneous trajectories, interactions between different elementswithin a frame of trajectory, which we call the ``Dimension-WiseInteractions'', would be more complex and challenging. However, most previousapproaches focus mainly on a specific form of trajectories, which means thesemethods could not be used to forecast heterogeneous trajectories, not tomention the dimension-wise interaction. Besides, previous methods mostly treattrajectory prediction as a normal time sequence generation task, indicatingthat these methods may require more work to directly analyze agents' behaviorsand social interactions at different temporal scales. In this paper, we bring anew ``view'' for trajectory prediction to model and forecast trajectorieshierarchically according to different frequency portions from the spectraldomain to learn to forecast trajectories by considering their frequencyresponses. Moreover, we try to expand the current trajectory prediction task byintroducing the dimension $M$ from ``another view'', thus extending itsapplication scenarios to heterogeneous trajectories vertically. Finally, weadopt the bilinear structure to fuse two factors, including the frequencyresponse and the dimension-wise interaction, to forecast heterogeneoustrajectories via spectrums hierarchically in a generic way. Experiments showthat the proposed model outperforms most state-of-the-art methods on ETH-UCY,Stanford Drone Dataset and nuScenes with heterogeneous trajectories, including2D coordinates, 2D and 3D bounding boxes.

Another Vertical View: A Hierarchical Network for Heterogeneous Trajectory Prediction via Spectrums