来自今天的爱可可AI前沿推介
[CV] Fully Differentiable RANSAC
T Wei, Y Patel, J Matas, D Barath
[Czech Technical University in Prague & ETH Zurich]
完全可微RANSAC
要点:
-
∇-RANSAC是一种完全可微的端到端可训练随机鲁棒估计器； -
∇-RANSAC在精度方面优于最先进的技术，是最快的方法之一； -
与特征匹配器一起训练，可以提高∇-RANSAC的精度。
摘要：
本文提出完全可微的∇ -RANSAC预测输入数据点的不可靠概率，利用引导采样器中的预测，并估计模型参数(如基本矩阵)及其质量，同时在整个过程中传播梯度。∇-RANSAC中的随机采样器基于一个精心设计的重参数化策略，即Gumbel Softmax采样器，允许将梯度直接传播到随后的可微最小求解器中。模型质量函数在∇-RANSAC内估计的所有模型的分数上被边缘化，以指导网络学习的精度和有用的概率。∇-RANSAC是第一个解锁几何估计管道端到端训练的方法，包括特征检测、匹配和类似RANSAC的随机鲁棒估计。为了证明其潜力，与LoFTR一起训练∇-RANSAC，即最近的无探测器特征匹配器，以端到端的方式找到可靠的对应关系。在一些基本和本质矩阵估计的现实世界数据集上测试了∇-RANSAC。在准确性方面，它优于最先进的，同时是最快的方法之一。
We propose the fully differentiable ∇ -RANSAC predicts the inlier probabilities of the input data points, exploits the predictions in a guided sampler, and estimates the model parameters (e.g., fundamental matrix) and its quality while propagating the gradients through the entire procedure. The random sampler in ∇-RANSAC is based on a clever re-parametrization strategy, i.e.\ the Gumbel Softmax sampler, that allows propagating the gradients directly into the subsequent differentiable minimal solver. The model quality function marginalizes over the scores from all models estimated within ∇-RANSAC to guide the network learning accurate and useful probabilities.∇-RANSAC is the first to unlock the end-to-end training of geometric estimation pipelines, containing feature detection, matching and RANSAC-like randomized robust estimation. As a proof of its potential, we train ∇-RANSAC together with LoFTR, i.e. a recent detector-free feature matcher, to find reliable correspondences in an end-to-end manner. We test ∇-RANSAC on a number of real-world datasets on fundamental and essential matrix estimation. It is superior to the state-of-the-art in terms of accuracy while being among the fastest methods. The code and trained models will be made public.
评论
沙发等你来抢