Generative Semantic Domain Adaptation for Perception in Autonomous Driving
Generative Semantic Domain Adaptation for Perception in Autonomous Driving
复制标题
DOI：
10.1007/s42421-022-00057-4
复制
发表时间：
2022-08
期刊：
影响因子：
--
通讯作者：
Amitangshu Mukherjee;Ameya Joshi;Anuj Sharma;C. Hegde;S. Sarkar
中科院分区：
文献类型：
--
作者：
Amitangshu Mukherjee;Ameya Joshi;Anuj Sharma;C. Hegde;S. Sarkar
自动驾驶系统依赖于感知和理解环境的能力来进行导航。神经网络是此类感知系统的构建块，训练这些网络需要大量不同的训练数据，其中包括地形、物体类别和不利照明/天气条件等不同类型的驾驶场景。然而，大多数公开可用的交通数据集都是在干净的天气和照明条件下采样的。数据增强通常被用作提高训练数据多样性的策略，用于训练基于机器学习的感知系统。然而，标准的增强技术（例如平移和翻转）有助于神经网络泛化简单的空间变换，并且需要更细致的技术来准确地应对新测试场景中的语义变化。我们提出了一种称为“语义域适应”的新数据增强方法，该方法依赖于属性条件生成模型的使用。我们通过分析深度网络在基于感知的任务中的表现，例如对 (i) 在一天中的不同时间和 (ii) 在不同天气条件下捕获的不同交通对象数据集进行分类和检测，并与使用传统增强方法训练的模型进行比较，证明这种数据增强提高了深度网络的泛化能力。我们进一步表明，基于 GAN 的增强分类模型比非基于 GAN 的增强模型更能抵御参数对抗攻击。
Autonomous driving systems depend on their ability to perceive and understand their environments for navigation. Neural networks are the building blocks of such perception systems, and training these networks requires vast amounts of diverse training data that includes different kinds of driving scenarios in terms of terrains, object categories, and adverse illumination/weather conditions. However, most publicly available traffic datasets suffer from having been sampled under clean weather and illumination conditions. Data augmentation is often used as a strategy to improve the diversity of training data for training machine learning-based perception systems. However, standard augmentation techniques (such as translation and flipping) help neural networks to generalize over simple spatial transformations and more nuanced techniques are required to accurately combat semantic variations in novel test scenarios. We propose a new data augmentation method called “semantic domain adaptation” that relies on the use of attribute-conditioned generative models. We show that such data augmentation improves the generalization capability of deep networks by analyzing their performance in perception-based tasks such as classification and detection on different datasets of traffic objects that are captured (i) at different times of the day and (ii) across different weather conditions, and comparing with models trained using traditional augmentation methods. We further show that GAN-based augmented classification models are more robust against parametric adversarial attacks than the non-GAN-based augmentation models.