论文阅读 [TPAMI-2022] Deep Visual Odometry With Adaptive Memory
论文阅读 [TPAMI-2022] Deep Visual Odometry With Adaptive Memory
论文搜索(studyai.com)
搜索论文: Deep Visual Odometry With Adaptive Memory
搜索论文: http://www.studyai.com/search/whole-site/?q=Deep+Visual+Odometry+With+Adaptive+Memory
关键字(Keywords)
Cameras; Task analysis; Tracking; Simultaneous localization and mapping; Pose estimation; History; Visual odometry; recurrent neural networks; memory; attention
机器学习; 机器视觉
姿态估计; 递归神经网络; 时间与空间; 视觉里程计; 相机姿态估计; RGBD数据
摘要(Abstract)
We propose a novel deep visual odometry (VO) method that considers global information by selecting memory and refining poses.
我们提出了一种新的深度视觉里程计（VO）方法，该方法通过选择记忆和细化姿势来考虑全局信息。.
Existing learning-based methods take the VO task as a pure tracking problem via recovering camera poses from image snippets, leading to severe error accumulation.
现有的基于学习的方法将VO任务视为一个纯跟踪问题，通过从图像片段中恢复相机姿态，导致严重的错误积累。.
Global information is crucial for alleviating accumulated errors.
全球信息对于减少累积错误至关重要。.
However, it is challenging to effectively preserve such information for end-to-end systems.
然而，为端到端系统有效保存此类信息是一个挑战。.
To deal with this challenge, we design an adaptive memory module, which progressively and adaptively saves the information from local to global in a neural analogue of memory, enabling our system to process long-term dependency.
为了应对这一挑战，我们设计了一个自适应存储模块，该模块将信息从局部到全局逐步自适应地保存在一个类似记忆的神经系统中，使我们的系统能够处理长期依赖性。.
Benefiting from global information in the memory, previous results are further refined by an additional refining module.
得益于内存中的全局信息，以前的结果通过额外的细化模块进一步细化。.
With the guidance of previous outputs, we adopt a spatial-temporal attention to select features for each view based on the co-visibility in feature domain.
在前面输出的指导下，我们基于特征域中的共可见性，采用时空注意为每个视图选择特征。.
Specifically, our architecture consisting of Tracking, Remembering and Refining modules works beyond tracking.
具体来说，我们的体系结构由跟踪、记忆和细化模块组成，其工作范围超出了跟踪。.
Experiments on the KITTI and TUM-RGBD datasets demonstrate that our approach outperforms state-of-the-art methods by large margins and produces competitive results against classic approaches in regular scenes.
在KITTI和TUM-RGBD数据集上的实验表明，我们的方法在很大程度上优于最先进的方法，并在常规场景中产生与经典方法相比的竞争结果。.
Moreover, our model achieves outstanding performance in challenging scenarios such as texture-less regions and abrupt motions, where classic algorithms tend to fail…
此外，我们的模型在无纹理区域和突然运动等具有挑战性的场景中取得了优异的性能，而经典算法往往会失败。。.
作者(Authors)
[‘Fei Xue’, ‘Xin Wang’, ‘Junqiu Wang’, ‘Hongbin Zha’]