Qi Liu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Qi Liu has a total of 15 co-authors. Here are the top co-authors by number of shared publications:
- Yuchao Dai
School of Electronics and Information, Northwestern Polytechnical University, Xian, China
2 shared publications - Bin Fan
School of Electronics and Information, Northwestern Polytechnical University, Xian, China
1 shared publication Hui Deng
School of Electronics and Information, Northwestern Polytechnical University, and Shaanxi Key Laboratory of Information Acquisition and Processing, Xian, China
1 shared publication- Jianhua He
University of Essex
1 shared publication Jiawei Shi
School of Electronics and Information, Northwestern Polytechnical University, and Shaanxi Key Laboratory of Information Acquisition and Processing, Xian, China
1 shared publicationKangyang Zhu
Northwestern Polytechnical University,School of Electronics and Information,Xian,China
1 shared publication
Latest Publications
- 22 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
Multi-event representation and multi-level fusion for robust RGB-event object tracking
- 22 February 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 339
Multi-event representation and multi-level fusion for robust RGB-event object tracking
- 19 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (8) , 8003-8010
LoopRefine: Deep Camera Pose Estimation With Loop Consistency
Recently, pose estimation under sparse views ($\leq 10$) has witnessed significant advances with the development of deep learning. Most existing methods directly regress the absolute poses, demonstrating leading performance on benchmarks. However, directly regressing the scaled poses using deep neural networks is inherently ill-posed, resulting in overfitted models that perform poorly on diverse scenarios. In contrast, we resort to the well-posed solutions from traditional Structure-from-Motion (SfM) pipelines and propose LoopRefine, a diffusion model that assumes known camera intrinsics and estimates pairwise normalized camera relative poses and utilizes triplet coplanar constraints to align the scale of camera poses. Like traditional SfM methods, LoopRefine incrementally constructs camera triplets, and the scale ambiguities are resolved by gradually recovering the scale of poses and connecting the pose graph. To further improve the pose estimation accuracy during inference, we explore pose compatibility by randomly chaining the loop transformations on the pose graph and organizing iterative loop consistency-based optimization. Extensive experiments demonstrate the superiority of our method, and the generalization performance on both object-centered datasets and scene datasets also proves the effectiveness of integrated geometric constraints.
- 19 June 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Robotics and Automation Letters
- Vol. 10 (8) , 8003-8010
LoopRefine: Deep Camera Pose Estimation With Loop Consistency
Recently, pose estimation under sparse views ($\leq 10$) has witnessed significant advances with the development of deep learning. Most existing methods directly regress the absolute poses, demonstrating leading performance on benchmarks. However, directly regressing the scaled poses using deep neural networks is inherently ill-posed, resulting in overfitted models that perform poorly on diverse scenarios. In contrast, we resort to the well-posed solutions from traditional Structure-from-Motion (SfM) pipelines and propose LoopRefine, a diffusion model that assumes known camera intrinsics and estimates pairwise normalized camera relative poses and utilizes triplet coplanar constraints to align the scale of camera poses. Like traditional SfM methods, LoopRefine incrementally constructs camera triplets, and the scale ambiguities are resolved by gradually recovering the scale of poses and connecting the pose graph. To further improve the pose estimation accuracy during inference, we explore pose compatibility by randomly chaining the loop transformations on the pose graph and organizing iterative loop consistency-based optimization. Extensive experiments demonstrate the superiority of our method, and the generalization performance on both object-centered datasets and scene datasets also proves the effectiveness of integrated geometric constraints.
- 20 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 191-196
An Air-Sea Identification and Detection Confrontation Evolutionary Game Model
An evolutionary game model of air-to-sea identification and detection confrontation is proposed in order to solve the problem of how to choose the attack and defense strategies in the identification and detection stages. First, the mathematical model of an air-to-sea attack strategy in identification and detection stages is established, and then the mathematical model of ship defense strategy in identification and detection stages is established. On this basis, the air-to-sea identification and detection revenue function is constructed, and then the corresponding evolutionary game model of air-to-sea identification and detection is established. An algorithm is proposed to solve the air-to-sea identification and detection confrontation evolutionary game model by replicating the dynamic equation. The resultant simulation shows that: For the jamming party and the anti-jamming party, faced the prerequisite of not knowing what jamming mode the other party will take, the anti-jamming party’s choice of frequency agility and the jamming party’s adoption of frequency-sweeping jamming can be the optimal strategies irrespective of the enemy’s choice. The simulation results verify the validity of the game model, which can provide better decision support for the commander’s rational decisions.
- 20 October 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 191-196
An Air-Sea Identification and Detection Confrontation Evolutionary Game Model
An evolutionary game model of air-to-sea identification and detection confrontation is proposed in order to solve the problem of how to choose the attack and defense strategies in the identification and detection stages. First, the mathematical model of an air-to-sea attack strategy in identification and detection stages is established, and then the mathematical model of ship defense strategy in identification and detection stages is established. On this basis, the air-to-sea identification and detection revenue function is constructed, and then the corresponding evolutionary game model of air-to-sea identification and detection is established. An algorithm is proposed to solve the air-to-sea identification and detection confrontation evolutionary game model by replicating the dynamic equation. The resultant simulation shows that: For the jamming party and the anti-jamming party, faced the prerequisite of not knowing what jamming mode the other party will take, the anti-jamming party’s choice of frequency agility and the jamming party’s adoption of frequency-sweeping jamming can be the optimal strategies irrespective of the enemy’s choice. The simulation results verify the validity of the game model, which can provide better decision support for the commander’s rational decisions.
- 1 July 2009
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1549-1552
The Anchored Instruction for programming courses: Cases, tasks and knowledge transfer
Programming is the nucleus foundation course in computer subjects, and also is considered generally as a hard learning and teaching course. With the introduction of anchored instruction in the course teaching, the study environment focusing on some true cases has been created. tasks with different degree of difficulties have been assigned to students for leading study. Cultivation of knowledge transfer ability for students has been highlighted in teaching process. Lecturer can mobilize the students' interest and enthusiasm in learning, reduce the difficulty of the curricula entry, stimulate learning motivation and guide self study. Practice has proved that the approach can quickly improve the students' practical programming capability to solve problems by using computer, which has significantly teaching effectiveness.
- 1 July 2009
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1549-1552
The Anchored Instruction for programming courses: Cases, tasks and knowledge transfer
Programming is the nucleus foundation course in computer subjects, and also is considered generally as a hard learning and teaching course. With the introduction of anchored instruction in the course teaching, the study environment focusing on some true cases has been created. tasks with different degree of difficulties have been assigned to students for leading study. Cultivation of knowledge transfer ability for students has been highlighted in teaching process. Lecturer can mobilize the students' interest and enthusiasm in learning, reduce the difficulty of the curricula entry, stimulate learning motivation and guide self study. Practice has proved that the approach can quickly improve the students' practical programming capability to solve problems by using computer, which has significantly teaching effectiveness.