Adversarial attacks pose serious challenges for deep neural network (DNN)-based analysis of various input signals. In the case of 3D point clouds, methods have been developed to identify points that play a key role in the network decision, and these become crucial in generating existing adversarial attacks. For example, a saliency map approach is a popular method for identifying adversarial drop points, whose removal would significantly impact the network decision. Generally, methods for identifying adversarial points rely on the deep model itself in order to determine which points are critically important for the model's decision. This paper aims to provide a novel viewpoint on this problem, in which adversarial points can be predicted independently of the model. To this end, we define 14 point cloud features and use multiple linear regression to examine whether these features can be used for model-free adversarial point prediction, and which combination of features is best suited for this purpose. Experiments show that a suitable combination of features is able to predict adversarial points of three different networks -- PointNet, PointNet++, and DGCNN -- significantly better than a random guess. The results also provide further insight into DNNs for point cloud analysis, by showing which features play key roles in their decision-making process.
翻译:以深神经网络为基础对各种输入信号进行深度神经网络(DNN)分析是严重的挑战。在3D点云的情况下,已经开发了方法,以确定在网络决定中起关键作用的点,这些点在形成现有的对抗性攻击中变得至关重要。例如,突出的地图方法是一种查明对抗性下降点的流行方法,其去除将大大影响网络决定。一般而言,确定对立点的方法依靠深度模型本身来确定哪些点对于模型的决定至关重要。本文件旨在就这一问题提供一种新颖的观点,其中对对立点可以独立预测。为此,我们界定了14点云特征,并使用多重线性回归来审查这些特征是否可用于无模式的对抗性点预测,以及哪些特征组合最适合于这一目的。实验表明,各种特征的适当组合能够预测三个不同网络的对立点 -- -- PointNet、pointNet++和DGCNN -- -- 比随机猜测的要好得多。结果还有助于进一步洞察 DNNP的云点分析,同时展示其关键作用。