In this paper, we propose HOME, a framework tackling the motion forecasting problem with an image output representing the probability distribution of the agent's future location. This method allows for a simple architecture with classic convolution networks coupled with attention mechanism for agent interactions, and outputs an unconstrained 2D top-view representation of the agent's possible future. Based on this output, we design two methods to sample a finite set of agent's future locations. These methods allow us to control the optimization trade-off between miss rate and final displacement error for multiple modalities without having to retrain any part of the model. We apply our method to the Argoverse Motion Forecasting Benchmark and achieve 1st place on the online leaderboard.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    HOME: Heatmap Output for future Motion Estimation


    Contributors:


    Publication date :

    2021-09-19


    Size :

    3025568 byte




    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    Vehicle Detection based on Deep Learning Heatmap Estimation

    Theodosc, Ruddy / Denis, Dieumet / Blanc, Christophe et al. | IEEE | 2019


    Reconfigurable avionics systems heatmap

    WONG JASON L | European Patent Office | 2022

    Free access

    Reconfigurable Avionics Systems Heatmap

    WONG JASON L | European Patent Office | 2021

    Free access

    Automated Motion Heatmap Generation for Bridge Navigation Watch Monitoring System

    Gokcek, Veysel / Kocak, Gazi / Genc, Yakup | Online Contents | 2022


    Automated Motion Heatmap Generation for Bridge Navigation Watch Monitoring System

    Gokcek Veysel / Kocak Gazi / Genc Yakup | DOAJ | 2022

    Free access