相关参考文献
注意：仅列出部分参考文献，下载原文获取全部文献信息。
Article
Engineering, Electrical & Electronic
Runze Li et al.
Summary: In this work, a novel framework called MonoIndoor++ is proposed to address the challenges in self-supervised monocular depth estimation for indoor environments. By introducing a depth factorization module, a residual pose estimation module, and coordinate convolutional encoding, the proposed method achieves state-of-the-art performance on benchmark indoor datasets.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Rene Ranftl et al.
Summary: The success of monocular depth estimation relies on large and diverse training sets. This study proposes tools and methods to mix different datasets and improve the performance of monocular depth estimation.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Weihao Yuan et al.
Summary: Estimating accurate depth from a single image is challenging, but this study proposes a CRFs optimization approach that leverages fully-connected CRFs and a multi-head attention mechanism to optimize the depth map. Experimental results show significant improvements over previous methods on multiple datasets, and the proposed method also outperforms existing panorama methods.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022)
(2022)
Article
Robotics
Carlos Campos et al.
Summary: ORB-SLAM3 is the first system capable of performing various SLAM tasks, including visual, visual-inertial, and multi-map SLAM, with improved accuracy and robustness, while also being able to survive in situations with poor visual information and achieve high accuracy.
IEEE TRANSACTIONS ON ROBOTICS
(2021)
Article
Robotics
Hualie Jiang et al.
Summary: Learning depth from spherical panoramas is a popular research topic, using a new framework to efficiently fuse features from equirectangular and cubemap projections only at the decoding stage to achieve state-of-the-art performance on popular datasets. Experiments verify the effectiveness of the proposed fusion strategy and module, showing advantages in model complexity and generalization capability.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Giovanni Pintore et al.
Summary: The study introduces a novel deep neural network for estimating depth maps from single monocular indoor panoramas. By compactly representing the scene into vertical slices of the sphere and exploiting relationships among slices, the network is able to recover the equirectangular depth map. Experimental results demonstrate that the method outperforms current solutions in prediction accuracy.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021
(2021)
Article
Engineering, Electrical & Electronic
Hong-Xiang Chen et al.
Summary: In this study, a Distortion-Aware Monocular Omnidirectional (DAMO) network was proposed to estimate dense depth maps from indoor panoramas. The network utilizes distortion-aware modules and a spherical-aware weight matrix to effectively extract features and alleviate bias caused by distortion. Experiments show state-of-the-art performance on the 360D dataset with high efficiency.
IEEE SIGNAL PROCESSING LETTERS
(2021)
Article
Computer Science, Software Engineering
Ana Serrano et al.
IEEE TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS
(2019)
Article
Computer Science, Software Engineering
P Pérez et al.
ACM TRANSACTIONS ON GRAPHICS
(2003)
