隐藏的足迹：从3D人类足迹中学习上下文步行能力
Hidden Footprints: Learning Contextual Walkability from 3D Human Trails
摘要：
预测人们可以在场景中行走的位置对于许多任务都很重要，包括自动驾驶系统和人类行为分析。然而，由于语义模糊和缺乏标记数据，为此目的学习计算模型是具有挑战性的：当前的数据集只告诉你人们在哪里，而不是他们可能在哪里。我们通过利用现有数据集的信息来解决这个问题，而不需要额外的标签。我们首先通过在图像之间传播人的观察来增加有效的、标记的可步行区域的集合，利用3D信息来创建我们所谓的隐藏足迹。然而，这种增强的数据仍然稀疏。我们设计了一个训练策略，设计这样的稀疏标签，结合类平衡的分类损失与上下文对抗损失。使用这种策略，我们展示了一个模型，学习从单个图像预测步行地图。我们在Waymo和Cityscapes数据集上评估了我们的模型，与基线和最先进的模型相比，表现出卓越的性能。
英文摘要：
Predicting where people can walk in a scene is important for many tasks, including autonomous driving systems and human behavior analysis. Yet learning a computational model for this purpose is challenging due to semantic ambiguity and a lack of labeled data: current datasets only tell you where people are, not where they could be. We tackle this problem by leveraging information from existing datasets, without additional labeling. We first augment the set of valid, labeled walkable regions by propagating person observations between images, utilizing 3D information to create what we call hidden footprints. However, this augmented data is still sparse. We devise a training strategy designed for such sparse labels, combining a class-balanced classification loss with a contextual adversarial loss. Using this strategy, we demonstrate a model that learns to predict a walkability map from a single image. We evaluate our model on the Waymo and Cityscapes datasets, demonstrating superior performance compared to baselines and state-of-the-art models.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
