Teqi Wang
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
Teqi Wang has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Dongdong Gui
Southeast University, Nanjing, China
2 shared publications - Juwei Guan
Southeast University, Nanjing, China
2 shared publications - Tongxin Zhu
School of Computer Science and Engineering, Southeast University, Nanjing, China
2 shared publications - Xiaolin Fang
School of Computer Science and Engineering, Southeast University, Nanjing, China
2 shared publications Daiying Zhou
School of Information and Communication Engineering, University of Electronic Science and Technology of China, Chengdu, China
1 shared publication- Guang Kou
Defense Innovation Institute, Beijing, China
1 shared publication
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 2519-2528
Imagine the Unseen: Vision-Language Anchored Synthesis for Open-Set Camouflaged Object Recognition
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 2519-2528
Open AccessImagine the Unseen: Vision-Language Anchored Synthesis for Open-Set Camouflaged Object Recognition
- Open Access
DRRNet: Macro-Micro Feature Fusion and Dual Reverse Refinement for Camouflaged Object Detection
The core challenge in Camouflage Object Detection (COD) lies in the indistinguishable similarity between targets and backgrounds in terms of color, texture, and shape. This causes existing methods to either lose edge details (such as hair-like fine structures) due to over-reliance on global semantic information or be disturbed by similar backgrounds (such as vegetation patterns) when relying solely on local features. We propose DRRNet, a four-stage architecture characterized by a "context-detail-fusion-refinement" pipeline to address these issues. Specifically, we introduce an Omni-Context Feature Extraction Module to capture global camouflage patterns and a Local Detail Extraction Module to supplement microstructural information for the full-scene context module. We then design a module for forming dual representations of scene understanding and structural awareness, which fuses panoramic features and local features across various scales. In the decoder, we also introduce a reverse refinement module that leverages spatial edge priors and frequency-domain noise suppression to perform a two-stage inverse refinement of the output. By applying two successive rounds of inverse refinement, the model effectively suppresses background interference and enhances the continuity of object boundaries. Experimental results demonstrate that DRRNet significantly outperforms state-of-the-art methods on benchmark datasets. Our code is available at https://github.com/jerrySunning/DRRNet.
Open AccessDRRNet: Macro-Micro Feature Fusion and Dual Reverse Refinement for Camouflaged Object Detection
The core challenge in Camouflage Object Detection (COD) lies in the indistinguishable similarity between targets and backgrounds in terms of color, texture, and shape. This causes existing methods to either lose edge details (such as hair-like fine structures) due to over-reliance on global semantic information or be disturbed by similar backgrounds (such as vegetation patterns) when relying solely on local features. We propose DRRNet, a four-stage architecture characterized by a "context-detail-fusion-refinement" pipeline to address these issues. Specifically, we introduce an Omni-Context Feature Extraction Module to capture global camouflage patterns and a Local Detail Extraction Module to supplement microstructural information for the full-scene context module. We then design a module for forming dual representations of scene understanding and structural awareness, which fuses panoramic features and local features across various scales. In the decoder, we also introduce a reverse refinement module that leverages spatial edge priors and frequency-domain noise suppression to perform a two-stage inverse refinement of the output. By applying two successive rounds of inverse refinement, the model effectively suppresses background interference and enhances the continuity of object boundaries. Experimental results demonstrate that DRRNet significantly outperforms state-of-the-art methods on benchmark datasets. Our code is available at https://github.com/jerrySunning/DRRNet.
- 10 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 59 (3) , 3129-3141
An Improved Method for Estimating Ballistic Target Micromotion and Structural Parameters Based on ISAR Image Sequences
Micro-motion and structural parameters extraction is a critical technique for recognizing ballistic missiles. It is generally known that parameter estimation methods based on inverse synthetic aperture radar(ISAR) image sequences can effectively extract target information. However, when the signal-to-noise ratio(SNR) is low, the traditional methods are prone to cross-association problems in the nearest-neighbor association procedure, and the structural parameters obtained may only be the local optimal solution. To solve the above two problems, an improved method for estimating ballistic target micro-motion and structural parameters based on ISAR image sequences is proposed in this paper. A distance association probability revision matrix is used to address the cross-association problem, reducing the risk of association with the incorrect nearest neighbor; at the same time, the impact of noise on the frequency estimation is decreased by utilizing refinement domain search-match algorithm, which can discover a narrower estimation interval with less noise to improve SNR. The global optimization capacity for structural parameters estimation of the traditional particle swarm optimization(PSO) algorithm is enhanced by step-by-step optimization algorithm, which reduces the dimension of the high-dimensional micro-motion projection cost function to ensure optimization direction variety. Simulation results show that when SNR is -5dB, the accuracy of the precession frequency estimation is improved by 0.2Hz compared with the traditional FFT method, and the RMSE values for cone height, precession angle, and nutation angle are 0.02m, 0.2 $°$ , and 0.3 $°$ , respectively, which are better than the traditional PSO algorithm.
- 10 November 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Aerospace and Electronic Systems
- Vol. 59 (3) , 3129-3141
An Improved Method for Estimating Ballistic Target Micromotion and Structural Parameters Based on ISAR Image Sequences
Micro-motion and structural parameters extraction is a critical technique for recognizing ballistic missiles. It is generally known that parameter estimation methods based on inverse synthetic aperture radar(ISAR) image sequences can effectively extract target information. However, when the signal-to-noise ratio(SNR) is low, the traditional methods are prone to cross-association problems in the nearest-neighbor association procedure, and the structural parameters obtained may only be the local optimal solution. To solve the above two problems, an improved method for estimating ballistic target micro-motion and structural parameters based on ISAR image sequences is proposed in this paper. A distance association probability revision matrix is used to address the cross-association problem, reducing the risk of association with the incorrect nearest neighbor; at the same time, the impact of noise on the frequency estimation is decreased by utilizing refinement domain search-match algorithm, which can discover a narrower estimation interval with less noise to improve SNR. The global optimization capacity for structural parameters estimation of the traditional particle swarm optimization(PSO) algorithm is enhanced by step-by-step optimization algorithm, which reduces the dimension of the high-dimensional micro-motion projection cost function to ensure optimization direction variety. Simulation results show that when SNR is -5dB, the accuracy of the precession frequency estimation is improved by 0.2Hz compared with the traditional FFT method, and the RMSE values for cone height, precession angle, and nutation angle are 0.02m, 0.2 $°$ , and 0.3 $°$ , respectively, which are better than the traditional PSO algorithm.