当前位置: X-MOL 学术Rob. Auton. Syst. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Visual place recognition using directed acyclic graph association measures and mutual information-based feature selection
Robotics and Autonomous Systems ( IF 4.3 ) Pub Date : 2020-10-01 , DOI: 10.1016/j.robot.2020.103598
Jurica Maltar , Ivan Marković , Ivan Petrović

Abstract Visual localization is a challenging problem, especially over the long run, since places can exhibit significant variation due to dynamic environmental and seasonal changes. To tackle this problem, we propose a visual place recognition method based on directed acyclic graph matching and feature maps extracted from deep convolutional neural networks (DCNN). Furthermore, in order to find the best subset of DCNN feature maps with minimal redundancy, we propose to form probability distributions on image representation features and leverage the Jensen–Shannon divergence to rank features. We evaluate the proposed approach on two challenging public datasets, namely the Bonn and the Freiburg datasets, and compare it to the state-of-the-art methods. For image representations, we evaluated the following DCNN architectures: AlexNet, OverFeat, ResNet18 and ResNet50. Due to the proposed graph structure, we are able to account for any kind of correlations in image sequences, and therefore dub our approach NOSeqSLAM. Algorithms with and without feature selection were evaluated based on precision–recall curves, area under the curve score, best recall at 100% precision score and running time, with NOSeqSLAM outperforming the counterpart approaches. Furthermore, by formulating the mutual information-based feature selection specifically for visual place recognition and by selecting the feature percentile with the best score, all the algorithms, and not just NOSeqSLAM, exhibited enhanced performance with the reduced feature set.

中文翻译:

使用有向无环图关联度量和基于互信息的特征选择的视觉位置识别

摘要 视觉定位是一个具有挑战性的问题,尤其是从长远来看,因为地方会因动态环境和季节变化而表现出显着的变化。为了解决这个问题,我们提出了一种基于有向无环图匹配和从深度卷积神经网络(DCNN)中提取的特征图的视觉位置识别方法。此外,为了找到冗余最小的 DCNN 特征图的最佳子集,我们建议在图像表示特征上形成概率分布,并利用 Jensen-Shannon 散度对特征进行排序。我们在两个具有挑战性的公共数据集(即波恩和弗莱堡数据集)上评估所提出的方法,并将其与最先进的方法进行比较。对于图像表示,我们评估了以下 DCNN 架构:AlexNet、OverFeat、ResNet18 和 ResNet50。由于提出的图结构,我们能够解释图像序列中的任何类型的相关性,因此将我们的方法称为 NOSeqSLAM。使用和不使用特征选择的算法根据精度-召回曲线、曲线下面积、100% 精度时的最佳召回率和运行时间进行评估,其中 NOSeqSLAM 优于对应方法。此外,通过专门为视觉位置识别制定基于互信息的特征选择,并通过选择得分最高的特征百分位数,所有算法,而不仅仅是 NOSeqSLAM,在减少特征集的情况下表现出增强的性能。使用和不使用特征选择的算法根据精度-召回曲线、曲线下面积、100% 精度时的最佳召回率和运行时间进行评估,其中 NOSeqSLAM 优于对应方法。此外,通过专门为视觉位置识别制定基于互信息的特征选择,并通过选择得分最高的特征百分位数,所有算法,而不仅仅是 NOSeqSLAM,在减少特征集的情况下表现出增强的性能。使用和不使用特征选择的算法根据精度-召回曲线、曲线下面积、100% 精度时的最佳召回率和运行时间进行评估,其中 NOSeqSLAM 优于对应方法。此外,通过专门为视觉位置识别制定基于互信息的特征选择,并通过选择得分最高的特征百分位数,所有算法,而不仅仅是 NOSeqSLAM,在减少特征集的情况下表现出增强的性能。
更新日期:2020-10-01
down
wechat
bug