Estimating the 3DoF rotation from a single RGB image is an important yet challenging problem. Probabilistic rotation regression has raised more and more attention with the benefit of expressing uncertainty information along with the prediction. Though modeling noise using Gaussian-resembling Bingham distribution and matrix Fisher distribution is natural, they are shown to be sensitive to outliers for the nature of quadratic punishment to deviations. In this paper, we draw inspiration from multivariate Laplace distribution and propose a novel Rotation Laplace distribution on SO(3). Rotation Laplace distribution is robust to the disturbance of outliers and enforces much gradient to the low-error region, resulting in a better convergence. Our extensive experiments show that our proposed distribution achieves state-of-the-art performance for rotation regression tasks over both probabilistic and non-probabilistic baselines. Our project page is at https://pku-epic.github.io/RotationLaplace.
翻译:从一个 RGB 图像中估算 3DoF 旋转是一个重要但具有挑战性的问题。 概率旋转回归已引起越来越多的关注, 并有利于在预测的同时表达不确定信息。 虽然使用高相比重比因汉姆分布和Fisher 矩阵分布的模型噪音是自然的, 但事实证明它们对于四边形惩罚的偏差性质非常敏感。 在本文中, 我们从多变量 Laplace 分布中得到灵感, 并提议在 SO(3) 上进行新的旋转拉普尔分布。 旋转拉普尔分布对于外端的干扰非常活跃, 并且将高梯度加到低梯度区域, 从而导致更好的趋同。 我们的广泛实验显示, 我们拟议的分布在概率和非概率基线上实现了最先进的旋转回归任务。 我们的项目网页是 https://pku-epic.github.io/RotationLapat。</s>