Abstract: This letter proposes an accurate pose refinement system that integrates point cloud registration and vehicle prior shapes to improve LiDAR-based vehicle pose estimation. For safe autonomous ...
Abstract: Deep learning models are highly susceptible to adversarial attacks, where subtle perturbations in the input images lead to misclassifications. Adversarial examples typically distort specific ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results