arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2005.06136cs.CV
收藏

在线自适应的自监督深度视觉里程计

Self-Supervised Deep Visual Odometry with Online Adaptation

Peking University(北京大学);PKU-SenseTime Machine Vision Joint Lab(北京大学-商汤科技机器视觉联合实验室)

Shunkai Li, Xin Wang, Yingdian Cao, Fei Xue, Zike Yan, Hongbin Zha

摘要:

自监督VO方法在从视频中联合估计相机姿态和深度方面取得了巨大成功。然而,与大多数数据驱动方法一样,现有的VO网络在面对与训练数据不同的场景时性能会显着下降,这使得它们不适合实际应用。在本文中,我们提出了一种在线Meta学习算法,使VO网络不断适应新的环境中的自我监督的方式。该方法利用卷积长短期记忆(convLSTM)来聚合过去丰富的时空信息。该网络能够记忆和学习其过去的经验,以更好地估计和快速适应当前帧。在开放环境下运行虚拟现实时,为了适应不断变化的环境,我们提出了一种在线特征对齐方法,通过对齐不同时刻的特征分布来实现。我们的VO网络能够无缝适应不同的环境。在看不见的室外场景,虚拟到现实世界和室外到室内环境中进行的大量实验表明,我们的方法始终优于最先进的自监督VO基线。

英文摘要:

Self-supervised VO methods have shown great success in jointly estimating camera pose and depth from videos. However, like most data-driven methods, existing VO networks suffer from a notable decrease in performance when confronted with scenes different from the training data, which makes them unsuitable for practical applications. In this paper, we propose an online meta-learning algorithm to enable VO networks to continuously adapt to new environments in a self-supervised manner. The proposed method utilizes convolutional long short-term memory (convLSTM) to aggregate rich spatial-temporal information in the past. The network is able to memorize and learn from its past experience for better estimation and fast adaptation to the current frame. When running VO in the open world, in order to deal with the changing environment, we propose an online feature alignment method by aligning feature distributions at different time. Our VO network is able to seamlessly adapt to different environments. Extensive experiments on unseen outdoor scenes, virtual to real world and outdoor to indoor environments demonstrate that our method consistently outperforms state-of-the-art self-supervised VO baselines considerably.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑