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

Predicting the future trajectory of a person remains a challenging problem,due to randomness and subjectivity of human movement. However, the movingpatterns of human in a constrained scenario typically conform to a limitednumber of regularities to a certain extent, because of the scenariorestrictions and person-person or person-object interactivity. Thus, anindividual person in this scenario should follow one of the regularities aswell. In other words, a person's subsequent trajectory has likely been traveledby others. Based on this hypothesis, we propose to forecast a person's futuretrajectory by learning from the implicit scene regularities. We call theregularities, inherently derived from the past dynamics of the people and theenvironment in the scene, scene history. We categorize scene historyinformation into two types: historical group trajectory andindividual-surroundings interaction. To exploit these two types of informationfor trajectory prediction, we propose a novel framework Scene HistoryExcavating Network (SHENet), where the scene history is leveraged in a simpleyet effective approach. In particular, we design two components: the grouptrajectory bank module to extract representative group trajectories as thecandidate for future path, and the cross-modal interaction module to model theinteraction between individual past trajectory and its surroundings fortrajectory refinement. In addition, to mitigate the uncertainty in ground-truthtrajectory, caused by the aforementioned randomness and subjectivity of humanmovement, we propose to include smoothness into the training process andevaluation metrics. We conduct extensive evaluations to validate the efficacyof our proposed framework on ETH, UCY, as well as a new, challenging benchmarkdataset PAV, demonstrating superior performance compared to state-of-the-artmethods.

Forecasting Human Trajectory from Scene History