HOME: Heatmap Output for future Motion Estimation

TL;DR

HOME introduces a novel heatmap output framework for motion forecasting, enabling flexible sampling of future agent locations and achieving top performance on the Argoverse benchmark.

Contribution

The paper presents a new heatmap-based approach for motion prediction that simplifies architecture and allows flexible trade-offs without retraining.

Findings

Achieved 1st place on Argoverse leaderboard.

Developed two sampling methods for future location prediction.

Demonstrated effective control over miss rate and displacement error.

Abstract

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.