|
作者 | Wan YM(万一鸣)1,2; Gao W(高伟)1,2; Han S(韩胜)1,2; Wu YH(吴毅红)1,2; Wu Yihong; Wan Yiming; Gao Wei
|
出版日期 | 2020-10
|
会议日期 | 2020.10.25-2020.10.28
|
会议地点 | Abu Dhabi, United Arab Emirates
|
英文摘要 | Visual localization is a fundamental problem in computer vision
and robotics. Recently, deep learning has shown to be effective
for robust monocular localization. Most deep learningbased
methods utilize convolution neural network (CNN) to
regress global 6 degree-of-freedom (Dof) pose. However,
these methods suffer from pose sparsity, leading to over-fitting
during training and poor localization performance on unseen
data. In this paper, we try to alleviate this issue by implementing
randomly geometric augmentation (RGA) during training.
Specifically, we firstly estimate the depth map using a depth
estimation network for the initial training image. Combing
the estimated depth, RGB image and its corresponding pose,
we can randomly synthesize new images of different views.
The synthesized and initial images are used to train the pose
regression network. Experiment results show our geometric
augmentation strategy can significantly improve the localization
accuracy. |
语种 | 英语
|
源URL | [http://ir.ia.ac.cn/handle/173211/39159] |
专题 | 自动化研究所_模式识别国家重点实验室_机器人视觉团队
|
作者单位 | 1.中国科学院大学 2.中国科学院自动化研究所
|
推荐引用方式 GB/T 7714 |
Wan YM,Gao W,Han S,et al. BOOSTING IMAGE-BASED LOCALIZATION VIA RANDOMLY GEOMETRIC DATA AUGMENTATION[C]. 见:. Abu Dhabi, United Arab Emirates. 2020.10.25-2020.10.28.
|