On the sensitivity of pose estimation neural networks: rotation parameterizations, Lipschitz constants, and provable bounds
This work addresses the problem of ensuring robustness in pose estimation for applications like robotics and computer vision, but it is incremental as it builds on existing Lipschitz constant methods and focuses on specific parameterizations.
The paper tackled the challenge of determining sensitivity bounds for pose estimation neural networks by developing a sensitivity measure based on Lipschitz constants and rotation parameterizations, showing that provable bounds are computable for networks using unconstrained exponential coordinates.
In this paper, we approach the task of determining sensitivity bounds for pose estimation neural networks. This task is particularly challenging as it requires characterizing the sensitivity of 3D rotations. We develop a sensitivity measure that describes the maximum rotational change in a network's output with respect to a Euclidean change in its input. We show that this measure is a type of Lipschitz constant, and that it is bounded by the product of a network's Euclidean Lipschitz constant and an intrinsic property of a rotation parameterization which we call the "distance ratio constant". We derive the distance ratio constant for several rotation parameterizations, and then discuss why the structure of most of these parameterizations makes it difficult to construct a pose estimation network with provable sensitivity bounds. However, we show that sensitivity bounds can be computed for networks which parameterize rotation using unconstrained exponential coordinates. We then construct and train such a network and compute sensitivity bounds for it.