Yiming Su
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
Research Interests
Yiming Su has a total of 26 co-authors. Here are the top co-authors by number of shared publications:
- Jiandong Tian
Shenyang Institute of Automation, Chinese Academy of Sciences
6 shared publications - Yang Lu
State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China
6 shared publications - Chunhui Hao
State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China
5 shared publications - Junchao Zhang
Central South University
3 shared publications - Zhen Zhang
State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China
3 shared publications Jiandong Tian
2 shared publications
Latest Publications
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 35 (10577149) , 2276-2291
Unsupervised Domain Adaptive Object Detection via Semantic Consistency and Compactness Learning
Unsupervised domain adaptive object detection methods enhance model robustness in the target domain without requiring target-domain annotations. Despite notable progress, existing methods face two major challenges: 1) insufficient and inefficient learning of holistic feature consistency due to cumbersome pixel-level style matching and semantic discrepancy elimination between domains as well as the overlooking of their collaborative effect, and 2) unreliable learning of category feature compactness caused by poor-quality target-domain samples, inaccurate pseudo-labels and noisy cross-domain contrast paradigms. To address these challenges, we propose a novel Semantic Consistency and Compactness Learning (SCCL) network. For consistency learning, we introduce a Visual Adaptation-guided Semantic Alignment (VSA) module that achieves style matching through simple feature adaptation and incorporates a novel adversarial-free self-supervised method for feature disentanglement. The collaboration between these two aspects enables sufficient and efficient consistency learning. For reliable compactness learning, we develop a plug-and-play Instance Center-Contrastive (ICC) head that, for the first time, comprehensively addresses all three potential causes of unreliable learning through three integrated innovations, concerning sample pseudo-label quality enhancement, reliable sample storage and updating, and a robust sample contrast paradigm. Besides, the mutual reinforcement effect of VSA and ICC simultaneously enhances feature transferability and discriminability. Extensive experiments across four UDA object detection benchmarks with two baselines show that SCCL achieves superior adaptability and robustness.
- 19 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 35 (10577149) , 2276-2291
Unsupervised Domain Adaptive Object Detection via Semantic Consistency and Compactness Learning
Unsupervised domain adaptive object detection methods enhance model robustness in the target domain without requiring target-domain annotations. Despite notable progress, existing methods face two major challenges: 1) insufficient and inefficient learning of holistic feature consistency due to cumbersome pixel-level style matching and semantic discrepancy elimination between domains as well as the overlooking of their collaborative effect, and 2) unreliable learning of category feature compactness caused by poor-quality target-domain samples, inaccurate pseudo-labels and noisy cross-domain contrast paradigms. To address these challenges, we propose a novel Semantic Consistency and Compactness Learning (SCCL) network. For consistency learning, we introduce a Visual Adaptation-guided Semantic Alignment (VSA) module that achieves style matching through simple feature adaptation and incorporates a novel adversarial-free self-supervised method for feature disentanglement. The collaboration between these two aspects enables sufficient and efficient consistency learning. For reliable compactness learning, we develop a plug-and-play Instance Center-Contrastive (ICC) head that, for the first time, comprehensively addresses all three potential causes of unreliable learning through three integrated innovations, concerning sample pseudo-label quality enhancement, reliable sample storage and updating, and a robust sample contrast paradigm. Besides, the mutual reinforcement effect of VSA and ICC simultaneously enhances feature transferability and discriminability. Extensive experiments across four UDA object detection benchmarks with two baselines show that SCCL achieves superior adaptability and robustness.
- 15 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 8679-8693
A Novel Dehazing Approach: Recovery of Color and Polarization Information Using Polarized Characteristics
Polarization provides valuable physical information, making it beneficial for various computer vision tasks. However, haze reduces both the color and polarization information of a scene. While existing single-image dehazing methods can restore color information, they are poor at recovering polarization information. Furthermore, current polarization-based dehazing approaches neglect the physical mechanisms of polarization degradation, resulting in inaccurate reconstruction of polarization information. In this paper, we propose a novel polarization dehazing algorithm, along with a polarization degradation model, to accurately recover both polarization and color information. First, we combine two key characteristics (the polarization achromatism prior and polarization attenuation prior) with the polarization degradation model to precisely reconstruct the scene's polarization. Then, we utilize the reconstructed polarization information to recover the color information of the scene. Finally, a multi-scale fusion optimization framework is introduced to further enhance the image quality. Our method shows excellent performance on both real-world indoor and outdoor polarized images, outperforming existing dehazing algorithms in both objective evaluation metrics and subjective visual assessment.
- 15 September 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 8679-8693
A Novel Dehazing Approach: Recovery of Color and Polarization Information Using Polarized Characteristics
Polarization provides valuable physical information, making it beneficial for various computer vision tasks. However, haze reduces both the color and polarization information of a scene. While existing single-image dehazing methods can restore color information, they are poor at recovering polarization information. Furthermore, current polarization-based dehazing approaches neglect the physical mechanisms of polarization degradation, resulting in inaccurate reconstruction of polarization information. In this paper, we propose a novel polarization dehazing algorithm, along with a polarization degradation model, to accurately recover both polarization and color information. First, we combine two key characteristics (the polarization achromatism prior and polarization attenuation prior) with the polarization degradation model to precisely reconstruct the scene's polarization. Then, we utilize the reconstructed polarization information to recover the color information of the scene. Finally, a multi-scale fusion optimization framework is introduced to further enhance the image quality. Our method shows excellent performance on both real-world indoor and outdoor polarized images, outperforming existing dehazing algorithms in both objective evaluation metrics and subjective visual assessment.
- 23 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP (15513203) , 1-12
LFUID: Light Field-Based Underwater Image Formation, Restoration, and Real-World Dataset
Underwater imaging in turbid environments presents significant challenges for industrial applications due to optical distortions that severely degrade image quality. This article addresses the fundamental information deficit in traditional single-image restoration methods by introducing the first comprehensive Light Field Underwater Image Dataset, comprising 1356 images captured across diverse environments with 1820× 720× 14× 14× 3 resolution. We develop a novel physics-based image formation model that extends scattering principles to the 4-D light field domain, incorporating water attenuation coefficients and scattering properties while accounting for light field camera characteristics. Our model features three key components: an ambient underwater optical constant, a phase function for angular scattering, and a pixel position modulation function for spatial variations in backscatter. Based on this model, we propose a restoration algorithm that leverages complementary information across subaperture views. Experimental results demonstrate that our approach significantly outperforms state-of-the-art underwater image enhancement techniques across multiple metrics and diverse underwater scenes, establishing a promising new direction for underwater imaging applications.
- 23 July 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Industrial Informatics
- Vol. PP (15513203) , 1-12
LFUID: Light Field-Based Underwater Image Formation, Restoration, and Real-World Dataset
Underwater imaging in turbid environments presents significant challenges for industrial applications due to optical distortions that severely degrade image quality. This article addresses the fundamental information deficit in traditional single-image restoration methods by introducing the first comprehensive Light Field Underwater Image Dataset, comprising 1356 images captured across diverse environments with 1820× 720× 14× 14× 3 resolution. We develop a novel physics-based image formation model that extends scattering principles to the 4-D light field domain, incorporating water attenuation coefficients and scattering properties while accounting for light field camera characteristics. Our model features three key components: an ambient underwater optical constant, a phase function for angular scattering, and a pixel position modulation function for spatial variations in backscatter. Based on this model, we propose a restoration algorithm that leverages complementary information across subaperture views. Experimental results demonstrate that our approach significantly outperforms state-of-the-art underwater image enhancement techniques across multiple metrics and diverse underwater scenes, establishing a promising new direction for underwater imaging applications.
- 19 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (7) , 11508-11520
Universal Polarization Demosaicking Using Newton Polynomial and Residual Interpolation
The development of division-of-focal-plane technology has made it possible to acquire full polarization information in a single shot, but its spatially alternating arrangement of polarizers suffers from partial loss of information. Polarization demosaicking methods (PDMs) are essential for recovering high-quality polarization information. Most existing PDMs can be applied only to a few specific monochrome polarization filter arrays (MPFAs) and rely on correspondingly strict assumptions. In this paper, we propose a universal PDM using the Newton polynomial and residual interpolation that can be applied to many different types of MPFAs. Our method first interpolates the intensity component S0 of the Stokes vector via the Newton polynomial and reconstructs the polarization channels using residual interpolation guided by S0. Specifically, S0 can be expressed as a linear combination of the polarization channels of any three angles, and our method selects the three angles in turn, derives the combination coefficients to obtain four weighted mosaic images, and then weighted averages them to obtain a universal weighted mosaic image, and uses Newton polynomial interpolation to derive the guide image S0. To validate our method, we extend the commonly used image quality metrics for measuring the accuracy of the recovered polarization information. Extensive experiments show that our method achieves competitive or better performance compared to other state-of-the-art methods on different types of MPFAs or color PFAs (CPFAs). The source code is available at https://github.com/ymingsu/UNRI.
- 19 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 25 (7) , 11508-11520
Universal Polarization Demosaicking Using Newton Polynomial and Residual Interpolation
The development of division-of-focal-plane technology has made it possible to acquire full polarization information in a single shot, but its spatially alternating arrangement of polarizers suffers from partial loss of information. Polarization demosaicking methods (PDMs) are essential for recovering high-quality polarization information. Most existing PDMs can be applied only to a few specific monochrome polarization filter arrays (MPFAs) and rely on correspondingly strict assumptions. In this paper, we propose a universal PDM using the Newton polynomial and residual interpolation that can be applied to many different types of MPFAs. Our method first interpolates the intensity component S0 of the Stokes vector via the Newton polynomial and reconstructs the polarization channels using residual interpolation guided by S0. Specifically, S0 can be expressed as a linear combination of the polarization channels of any three angles, and our method selects the three angles in turn, derives the combination coefficients to obtain four weighted mosaic images, and then weighted averages them to obtain a universal weighted mosaic image, and uses Newton polynomial interpolation to derive the guide image S0. To validate our method, we extend the commonly used image quality metrics for measuring the accuracy of the recovered polarization information. Extensive experiments show that our method achieves competitive or better performance compared to other state-of-the-art methods on different types of MPFAs or color PFAs (CPFAs). The source code is available at https://github.com/ymingsu/UNRI.
- 15 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Imaging
- Vol. 10 (25730436) , 1400-1413
A Hybrid Polarization Image Demosaicking Algorithm Based on Inter-Channel Correlation
Emerging monochrome and chromatic polarization filter array (MPFA and CPFA) cameras require polarization demosaicking to obtain accurate polarization parameters. Polarization cameras sample the polarization intensity at each location of the pixels. A captured raw image must be converted to a full-channel polarization intensity image using the polarization demosaicking method (PDM). However, due to sparse sampling between polarization channels, implementing MPFA and CPFA demosaicking has been challenging. This paper proposes a new hybrid polarization demosaicking algorithm that leverages polarization confidence-based refinement to exploit inter-channel polarization correlation. Additionally, we enhance texture correlation to utilize inter-channel texture correlation fully. Our three-stage PDM preserves both the polarization and texture information. We also introduce a metric computation method to handle the $\pi$ -ambiguity of the angle of line polarization (AoLP). This approach mitigates inaccuracies and $\pi$ -ambiguity in existing methods when describing the quality of AoLP reconstruction. We extensively compare and conduct ablation experiments on synthetic datasets from MPFA and CPFA. Our method achieves competitive results compared to other state-of-the-art methods. Furthermore, we evaluate our proposal on real-world datasets to demonstrate its applicability in real-world, variable scenarios. Two application experiments (road detection and shape from polarization) show that our proposal can be applied to real-world applications.
- 15 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Computational Imaging
- Vol. 10 (25730436) , 1400-1413
A Hybrid Polarization Image Demosaicking Algorithm Based on Inter-Channel Correlation
Emerging monochrome and chromatic polarization filter array (MPFA and CPFA) cameras require polarization demosaicking to obtain accurate polarization parameters. Polarization cameras sample the polarization intensity at each location of the pixels. A captured raw image must be converted to a full-channel polarization intensity image using the polarization demosaicking method (PDM). However, due to sparse sampling between polarization channels, implementing MPFA and CPFA demosaicking has been challenging. This paper proposes a new hybrid polarization demosaicking algorithm that leverages polarization confidence-based refinement to exploit inter-channel polarization correlation. Additionally, we enhance texture correlation to utilize inter-channel texture correlation fully. Our three-stage PDM preserves both the polarization and texture information. We also introduce a metric computation method to handle the $\pi$ -ambiguity of the angle of line polarization (AoLP). This approach mitigates inaccuracies and $\pi$ -ambiguity in existing methods when describing the quality of AoLP reconstruction. We extensively compare and conduct ablation experiments on synthetic datasets from MPFA and CPFA. Our method achieves competitive results compared to other state-of-the-art methods. Furthermore, we evaluate our proposal on real-world datasets to demonstrate its applicability in real-world, variable scenarios. Two application experiments (road detection and shape from polarization) show that our proposal can be applied to real-world applications.
- 8 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 10 (3) , 1944-1958
Polarization-Aware Low-Light Image Enhancement for Nighttime Intelligent Vehicles
Polarization-based environmental perception is essential for intelligent vehicles. However, images taken in low-light conditions often have poor visibility. Deterioration of visual quality will decrease the accuracy of intelligent vehicle applications based on polarization, resulting in a decreased perception range and increased navigation risk. To suppress the effects of real-world nighttime scenes for intelligent vehicles, we propose a retinex-based polarization-aware low-light image enhancement network (RPLENet) that can enhance the polarization and color information. We prove that the reflectance layer obtained from the retinex decomposition of the unpolarized component $I_{un}$ is more accurate than the reflectance layer from any polarizer angle intensity image. Therefore, we perform retinex decomposition on $I_{un}$ to obtain a more precise reflectance layer instead of decomposing the reflectance layer in the polarizer angle intensity image. Our method achieves state-of-the-art performance by conducting extensive experiments on both synthetic and real-world datasets. We also collected polarization datasets from real-world nighttime road scenes to evaluate the performance of RPLENet. Our method can effectively improve the perception of polarization-based features in low-light conditions. This is of great value for the environmental robustness of visual perception systems and transportation safety. Our codes are available at https://github.com/authorsdesk/RPLENet .
- 8 August 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 10 (3) , 1944-1958
Polarization-Aware Low-Light Image Enhancement for Nighttime Intelligent Vehicles
Polarization-based environmental perception is essential for intelligent vehicles. However, images taken in low-light conditions often have poor visibility. Deterioration of visual quality will decrease the accuracy of intelligent vehicle applications based on polarization, resulting in a decreased perception range and increased navigation risk. To suppress the effects of real-world nighttime scenes for intelligent vehicles, we propose a retinex-based polarization-aware low-light image enhancement network (RPLENet) that can enhance the polarization and color information. We prove that the reflectance layer obtained from the retinex decomposition of the unpolarized component $I_{un}$ is more accurate than the reflectance layer from any polarizer angle intensity image. Therefore, we perform retinex decomposition on $I_{un}$ to obtain a more precise reflectance layer instead of decomposing the reflectance layer in the polarizer angle intensity image. Our method achieves state-of-the-art performance by conducting extensive experiments on both synthetic and real-world datasets. We also collected polarization datasets from real-world nighttime road scenes to evaluate the performance of RPLENet. Our method can effectively improve the perception of polarization-based features in low-light conditions. This is of great value for the environmental robustness of visual perception systems and transportation safety. Our codes are available at https://github.com/authorsdesk/RPLENet .
- 3 May 2024
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 179
Underwater descattering with polarization line constraint
- 3 May 2024
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 179
Underwater descattering with polarization line constraint
- 21 March 2024
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 178
Polarization image demosaicking based on homogeneity space
- 21 March 2024
- Published by Elsevier in Optics and Lasers in Engineering
- Vol. 178
Polarization image demosaicking based on homogeneity space
- 23 August 2017
- Published by Springer Nature
- p. 652-662
Research on Speed Identification of Induction Motor Based on Sliding Mode Observer
- 23 August 2017
- Published by Springer Nature
- p. 652-662
Research on Speed Identification of Induction Motor Based on Sliding Mode Observer