How to improve CNN-based 6-DoF camera pose estimation
This work addresses incremental improvements in camera pose estimation for robotics or AR applications.
The paper tackled the problem of improving CNN-based 6-DoF camera pose estimation by modifying PoseNet, focusing on field-of-view importance, data augmentation, and LSTM effects, resulting in enhanced performance for monocular pose regression.
Convolutional neural networks (CNNs) and transfer learning have recently been used for 6 degrees of freedom (6-DoF) camera pose estimation. While they do not reach the same accuracy as visual SLAM-based approaches and are restricted to a specific environment, they excel in robustness and can be applied even to a single image. In this paper, we study PoseNet [1] and investigate modifications based on datasets' characteristics to improve the accuracy of the pose estimates. In particular, we emphasize the importance of field-of-view over image resolution; we present a data augmentation scheme to reduce overfitting; we study the effect of Long-Short-Term-Memory (LSTM) cells. Lastly, we combine these modifications and improve PoseNet's performance for monocular CNN based camera pose regression.