no code implementations • ECCV 2020 • Huikun Bi, Ruisi Zhang, Tianlu Mao, Zhigang Deng, Zhaoqi Wang
This work presents a novel First-person View based Trajectory predicting model (FvTraj) to estimate the future trajectories of pedestrians in a scene given their observed trajectories and the corresponding first-person view images.
no code implementations • 11 Sep 2023 • Pengfei Yao, Tianlu Mao, Min Shi, Jingkai Sun, Zhaoqi Wang
We introduce expert attention, which adjusts the weights of different depths of network layers, avoiding the model updated slowly due to gradient problem and enabling fast learning of new scenario's knowledge to restore prediction accuracy.
no code implementations • 17 Aug 2023 • Zijian Song, Huikun Bi, Ruisi Zhang, Tianlu Mao, Zhaoqi Wang
We presented a cross-view trajectory prediction method using shared 3D Queries (XVTP3D).