Deep Projective Rotation Estimation through Relative Supervision
Deep Projective Rotation Estimation through Relative Supervision
复制标题
DOI：
10.48550/arxiv.2211.11182
复制
发表时间：
2022-11
期刊：
影响因子：
--
通讯作者：
Brian Okorn;Chuer Pan;M. Hebert;David Held
中科院分区：
文献类型：
--
作者：
Brian Okorn;Chuer Pan;M. Hebert;David Held
方向估计是各种视觉和机器人任务的核心，例如相机和物体姿态估计。深度学习提供了一种开发基于图像的方向估计器的方法;然而，这种估计器通常需要在大型标记数据集上进行训练，这可能是时间密集型的收集。在这项工作中，我们探索是否可以使用来自未标记数据的自监督学习来缓解这个问题。具体而言，我们假设访问相邻姿态之间的相对取向的估计，使得可以经由局部对准方法来获得。虽然自监督学习已成功用于平移对象关键点，但在这项工作中，我们表明，由于旋转空间的非凸性，天真地将相对监督应用于旋转群SO（3）$通常无法收敛。为了解决这一挑战，我们提出了一种新的自监督方向估计算法，该算法利用修改的罗德里格斯参数将$SO（3）$的闭流形立体投影到$\mathbb{R}^{3}$的开流形，允许在开放的欧几里得空间中进行优化。我们在两种设置中经验性地验证了所提出的旋转平均问题算法的好处：（1）直接优化旋转参数，以及（2）优化从图像预测对象方向的卷积神经网络的参数。在这两种设置中，我们证明我们提出的算法能够比纯粹在$SO（3）$空间中操作的算法更快地收敛到一致的相对方向框架。更多信息可以在https://sites.google.com/view/deep-projective-rotation/home上找到。
Orientation estimation is the core to a variety of vision and robotics tasks such as camera and object pose estimation. Deep learning has offered a way to develop image-based orientation estimators; however, such estimators often require training on a large labeled dataset, which can be time-intensive to collect. In this work, we explore whether self-supervised learning from unlabeled data can be used to alleviate this issue. Specifically, we assume access to estimates of the relative orientation between neighboring poses, such that can be obtained via a local alignment method. While self-supervised learning has been used successfully for translational object keypoints, in this work, we show that naively applying relative supervision to the rotational group $SO(3)$ will often fail to converge due to the non-convexity of the rotational space. To tackle this challenge, we propose a new algorithm for self-supervised orientation estimation which utilizes Modified Rodrigues Parameters to stereographically project the closed manifold of $SO(3)$ to the open manifold of $\mathbb{R}^{3}$, allowing the optimization to be done in an open Euclidean space. We empirically validate the benefits of the proposed algorithm for rotational averaging problem in two settings: (1) direct optimization on rotation parameters, and (2) optimization of parameters of a convolutional neural network that predicts object orientations from images. In both settings, we demonstrate that our proposed algorithm is able to converge to a consistent relative orientation frame much faster than algorithms that purely operate in the $SO(3)$ space. Additional information can be found at https://sites.google.com/view/deep-projective-rotation/home .