Interpretable Social Anchors for Human Trajectory Forecasting in Crowds

Parth Kothari, Brian Sifringer, Alexandre Alahi; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2021, pp. 15556-15566

Abstract


Human trajectory forecasting in crowds, at its core, is a sequence prediction problem with specific challenges of capturing inter-sequence dependencies (social interactions) and consequently predicting socially-compliant multimodal distributions. In recent years, neural network-based methods have been shown to outperform hand-crafted methods on distance-based metrics. However, these data-driven methods still suffer from one crucial limitation: lack of interpretability. To overcome this limitation, we leverage the power of discrete choice models to learn interpretable rule-based intents, and subsequently utilise the expressibility of neural networks to model scene-specific residual. Extensive experimentation on the interaction-centric benchmark TrajNet++ demonstrates the effectiveness of our proposed architecture to explain its predictions without compromising the accuracy.

Related Material


[pdf] [arXiv]
[bibtex]
@InProceedings{Kothari_2021_CVPR, author = {Kothari, Parth and Sifringer, Brian and Alahi, Alexandre}, title = {Interpretable Social Anchors for Human Trajectory Forecasting in Crowds}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, month = {June}, year = {2021}, pages = {15556-15566} }