Related references
Note: Only part of the references are listed.
Article
Robotics
Borna Besic et al.
Summary: Scene understanding is crucial for autonomous vehicles to navigate safely. This paper proposes AdaptLPS, a novel unsupervised domain adaptation approach for LiDAR panoptic segmentation. By combining data-based and model-based adaptations, AdaptLPS effectively reduces the domain gap between different LiDAR setups and achieves superior performance on real-world autonomous driving datasets.
IEEE ROBOTICS AND AUTOMATION LETTERS
(2022)
Article
Robotics
Daniele Cattaneo et al.
Summary: Loop closure detection is a crucial part of SLAM systems, and this article introduces a novel network called LCDNet, which effectively detects loop closures in LiDAR point clouds. It outperforms state-of-the-art techniques, especially for reverse loops, and shows good generalization ability across different sensor setups.
IEEE TRANSACTIONS ON ROBOTICS
(2022)
Article
Computer Science, Artificial Intelligence
Borna Besic et al.
Summary: This research addresses the problem of robot's perception affected by dynamic objects by synthesizing color, texture, and geometry in the occluded regions. The proposed architecture, DynaFill, incorporates a gated recurrent feedback mechanism and optimizes the synthesis process using adversarial training. It achieves spatially and temporally coherent results and is capable of correcting regions correlated with dynamic objects. The research also introduces a large-scale dataset and demonstrates the utility of the synthesized images in visual localization.
IEEE TRANSACTIONS ON INTELLIGENT VEHICLES
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Saisubramaniam Gopalakrishnan et al.
Summary: This article introduces flashcards, a method that captures and consolidates the knowledge representation of a neural network in a continual learning scenario to prevent catastrophic forgetting. Flashcards, as a replay strategy, effectively capture and substitute the knowledge representation of the original data, and have been empirically validated on multiple continual learning tasks.
2022 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION (WACV 2022)
(2022)
Proceedings Paper
Robotics
Mayank Mittal et al.
Summary: This paper introduces a life-saving technology using unmanned aerial vehicles equipped with bioradars to identify survivors after natural disasters. The technology requires UAVs to autonomously navigate and land on debris piles. The paper proposes a new landing site detection algorithm and conducts experiments using a synthetic dataset and a simulation environment.
ROBOTICS RESEARCH: THE 19TH INTERNATIONAL SYMPOSIUM ISRR
(2022)
Article
Automation & Control Systems
Ruihao Li et al.
Summary: DeepSLAM is an unsupervised deep learning-based visual SLAM system that uses stereo imagery for training and monocular image sequences for testing. It consists of essential components such as Mapping-Net, Tracking-Net, and Loop-Net, which enable it to generate pose estimates, depth maps, and outlier rejection masks simultaneously. The system demonstrates good performance in pose estimation accuracy and robustness in challenging scenes, as evaluated on various datasets.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2021)
Proceedings Paper
Automation & Control Systems
Gerhard Kurz et al.
Summary: The paper introduces a lifelong SLAM algorithm that addresses the speed issue in traditional SLAM methods through graph pruning while maintaining better accuracy. By using geometric criteria for vertex selection and marginalization method for robustness to wrong loop closures, the proposed algorithm achieves high speed and good accuracy even on long datasets.
2021 IEEE/RSJ INTERNATIONAL CONFERENCE ON INTELLIGENT ROBOTS AND SYSTEMS (IROS)
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Shunkai Li et al.
Summary: The research introduces an online adaptation framework for deep VO, leveraging geometric computations and Bayesian inference to enhance performance. This method not only addresses pose estimation but also solves the problem through optical flow and depth, continuously improving and optimizing for fast adaptation to unknown environments. Extensive experiments demonstrate that the proposed method achieves state-of-the-art generalization ability among self-supervised VO methods.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Yevhen Kuznietsov et al.
Summary: This paper proposes a novel self-supervised Continuous Monocular Depth Adaptation method (CoMoDA) that dynamically adjusts the pretrained model in test videos. By utilizing past experiences and preventing model overfitting, continuous adaptation to the same scene is achieved, leading to state-of-the-art depth estimation performance.
2021 IEEE WINTER CONFERENCE ON APPLICATIONS OF COMPUTER VISION WACV 2021
(2021)
Article
Robotics
Juana Valeria Hurtado et al.
Summary: The rapid advancements in robotics and machine learning have enabled robots to perform diverse tasks in everyday environments. A framework is proposed to reduce bias in robot navigation models, promoting fair human-robot interactions and positively impacting society.
FRONTIERS IN ROBOTICS AND AI
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Zhenyu Zhang et al.
2020 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR)
(2020)
Article
Computer Science, Information Systems
Hongcheng Luo et al.
IEEE TRANSACTIONS ON MULTIMEDIA
(2019)
Article
Computer Science, Artificial Intelligence
Zhizhong Li et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2018)
Article
Robotics
Will Maddern et al.
INTERNATIONAL JOURNAL OF ROBOTICS RESEARCH
(2017)
Article
Robotics
Raul Mur-Artal et al.
IEEE TRANSACTIONS ON ROBOTICS
(2015)
Article
Computer Science, Artificial Intelligence
Henrik Kretzschmar et al.
KUNSTLICHE INTELLIGENZ
(2010)