Explainable Action Prediction through Self-Supervision on Scene Graphs
Explainable Action Prediction through Self-Supervision on Scene Graphs
复制标题
DOI：
10.1109/icra48891.2023.10161132
复制
发表时间：
2023-02
期刊：
影响因子：
--
通讯作者：
Pawit Kochakarn;D. Martini;Daniel Omeiza;L. Kunze
中科院分区：
文献类型：
--
作者：
Pawit Kochakarn;D. Martini;Daniel Omeiza;L. Kunze
这项工作探索了场景图作为自动驾驶的高级信息的提取表示，并应用于未来的驾驶员行为预测。考虑到数据样本的稀缺性和强不平衡性，我们提出了一种自监督管道来推断具有代表性和良好分离的嵌入。关键方面是可解释性和可解释性；因此，我们在我们的体系结构中嵌入了可以在场景图上创建空间和时间热图的注意机制。我们根据完全监督的方法在道路数据集上评估我们的系统，显示了我们的训练制度的优越性。
This work explores scene graphs as a distilled representation of high-level information for autonomous driving, applied to future driver-action prediction. Given the scarcity and strong imbalance of data samples, we propose a self-supervision pipeline to infer representative and well-separated embeddings. Key aspects are interpretability and explainability; as such, we embed in our architecture attention mechanisms that can create spatial and temporal heatmaps on the scene graphs. We evaluate our system on the ROAD dataset against a fully-supervised approach, showing the superiority of our training regime.