Zihan Zhou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Zihan Zhou has a total of 37 co-authors. Here are the top co-authors by number of shared publications:
- Patrick Le Callet
Nantes Université, Nantes, France
4 shared publications - Yong Xu
School of Computer Science & Engineering, South China University of Technology, Guangzhou, China
4 shared publications Jing Li
3 shared publications- Jing Li
Alibaba Group, Beijing, China
3 shared publications - Ruotao Xu
School of Computer Science and Engineering, South China University of Technology, Guangzhou, China
3 shared publications Yong Xu
3 shared publications
Latest Publications
- 15 May 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 346
ObjectDiff: An object-centric diffusion policy with modality-specific conditioning for robot manipulation
- 15 May 2026
- Published by Elsevier in Knowledge-Based Systems
- Vol. 346
ObjectDiff: An object-centric diffusion policy with modality-specific conditioning for robot manipulation
- 2 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (4) , 4622-4635
Deep Underwater Image Quality Assessment via Progressive Physics-Aware Multi-Prior Collaboration
Underwater image quality assessment (UIQA) is a critical research area, challenged by underwater environments such as wavelength-dependent light attenuation, scattering, and non-uniform illumination. Existing deep learning-based UIQA methods often address these degradations in isolation, neglecting their complex interplay with human perception and lacking explicit modeling of underwater optical phenomena. To address this, we propose PhysIQ-Net, a novel framework that integrates physics-driven principles with progressive multi-prior interaction modeling through three key innovations: First, introduce dual physics-based decomposition that separates images into Backscatter, Transmission, Reflectance, and Illuminance components to capture distinct degradation mechanisms; Second, propose prior-guided dynamic filtering that adapts convolutional kernels to image-specific content using physical priors; and Third, propose physic-informed Cross-Domain Feature Interaction that enables bidirectional collaboration between color-aware and structure-aware representations to model their perceptual inter-dependencies. Extensive experiments across multiple benchmark datasets demonstrate that PhysIQ-Net significantly outperforms existing methods, with ablation studies validating each component’s contribution, providing a robust solution for UIQA.
- 2 December 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Circuits and Systems for Video Technology
- Vol. 36 (4) , 4622-4635
Deep Underwater Image Quality Assessment via Progressive Physics-Aware Multi-Prior Collaboration
Underwater image quality assessment (UIQA) is a critical research area, challenged by underwater environments such as wavelength-dependent light attenuation, scattering, and non-uniform illumination. Existing deep learning-based UIQA methods often address these degradations in isolation, neglecting their complex interplay with human perception and lacking explicit modeling of underwater optical phenomena. To address this, we propose PhysIQ-Net, a novel framework that integrates physics-driven principles with progressive multi-prior interaction modeling through three key innovations: First, introduce dual physics-based decomposition that separates images into Backscatter, Transmission, Reflectance, and Illuminance components to capture distinct degradation mechanisms; Second, propose prior-guided dynamic filtering that adapts convolutional kernels to image-specific content using physical priors; and Third, propose physic-informed Cross-Domain Feature Interaction that enables bidirectional collaboration between color-aware and structure-aware representations to model their perceptual inter-dependencies. Extensive experiments across multiple benchmark datasets demonstrate that PhysIQ-Net significantly outperforms existing methods, with ablation studies validating each component’s contribution, providing a robust solution for UIQA.
- 7 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 9712-9722
Enhancing CNN-Based Blind Image Quality Assessment via Deep Cross-Layer Pattern Encoding
Evaluating image quality without reference images, known as blind image quality assessment (BIQA), is crucial for image communication. Recently, convolutional neural networks (CNNs) have emerged as a prominent BIQA approach due to their feature learning power. Usually, both high-level semantic information and low-level details significantly impact perceived visual quality. However, most existing CNN-based methods focus on high-level semantic information via aggregating features on top of the last convolutional layer into a global descriptor, neglecting the importance of shallow, low-level cues. To address this limitation, this paper proposes a novel approach that exploits local encoding and histogram-based pyramid pooling on crosslayer features produced by a CNN, achieving a joint local and global analysis. Specifically, we introduce a cross-layer pattern encoding model that characterizes features generated along convolutional layers via a soft histogram of local 3D binary patterns. This leads to a highly informative yet compact descriptor for score regression. By building this module into a ResNet backbone, we present an effective BIQA model demonstrating state-ofthe-art performance in extensive experiments on synthetic and authentic datasets.
- 7 October 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 27 (15209210) , 9712-9722
Enhancing CNN-Based Blind Image Quality Assessment via Deep Cross-Layer Pattern Encoding
Evaluating image quality without reference images, known as blind image quality assessment (BIQA), is crucial for image communication. Recently, convolutional neural networks (CNNs) have emerged as a prominent BIQA approach due to their feature learning power. Usually, both high-level semantic information and low-level details significantly impact perceived visual quality. However, most existing CNN-based methods focus on high-level semantic information via aggregating features on top of the last convolutional layer into a global descriptor, neglecting the importance of shallow, low-level cues. To address this limitation, this paper proposes a novel approach that exploits local encoding and histogram-based pyramid pooling on crosslayer features produced by a CNN, achieving a joint local and global analysis. Specifically, we introduce a cross-layer pattern encoding model that characterizes features generated along convolutional layers via a soft histogram of local 3D binary patterns. This leads to a highly informative yet compact descriptor for score regression. By building this module into a ResNet backbone, we present an effective BIQA model demonstrating state-ofthe-art performance in extensive experiments on synthetic and authentic datasets.
- 30 May 2025
- Published by Elsevier in Pattern Recognition
- Vol. 169
Nighttime image dehazing via a physics-aware dynamic neural model with progressive contrastive regularization
- 30 May 2025
- Published by Elsevier in Pattern Recognition
- Vol. 169
Nighttime image dehazing via a physics-aware dynamic neural model with progressive contrastive regularization
- 17 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-16
Underwater Image Enhancement via Adaptive Bi-Level Color-Based Adjustment
Underwater images often exhibit severe color distortions and reduced contrast due to light absorption and scattering, presenting substantial challenges for image enhancement techniques. To address these challenges, this paper presents BCTA-Net, an adaptive bi-level color-based network specifically engineered to enhance the quality of underwater images by addressing distortions in dynamic and complex environments. The network integrates content-aware global and local restoration strategies. On a local scale, a color-aware attention mechanism is proposed which employs color histograms to adaptively correct non-uniform color distortions and enhance local color fidelity. Additionally, a triple attention module restores spatially varying local details in a content-aware manner, improving clarity and texture precision of enhancement. These elements are combined into a dual-branch architecture aimed at reducing local contrast, color fidelity, and detail precision issues. On a global scale, contrastive learning focused on background lightness corrects color distortions due to uneven illumination. The integration of these components results in a lightweight, dynamic global-local model with robust generalization capabilities across various underwater scenarios, as demonstrated by comprehensive experiments that show significant performance improvements over existing methods.
- 17 March 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-16
Underwater Image Enhancement via Adaptive Bi-Level Color-Based Adjustment
Underwater images often exhibit severe color distortions and reduced contrast due to light absorption and scattering, presenting substantial challenges for image enhancement techniques. To address these challenges, this paper presents BCTA-Net, an adaptive bi-level color-based network specifically engineered to enhance the quality of underwater images by addressing distortions in dynamic and complex environments. The network integrates content-aware global and local restoration strategies. On a local scale, a color-aware attention mechanism is proposed which employs color histograms to adaptively correct non-uniform color distortions and enhance local color fidelity. Additionally, a triple attention module restores spatially varying local details in a content-aware manner, improving clarity and texture precision of enhancement. These elements are combined into a dual-branch architecture aimed at reducing local contrast, color fidelity, and detail precision issues. On a global scale, contrastive learning focused on background lightness corrects color distortions due to uneven illumination. The integration of these components results in a lightweight, dynamic global-local model with robust generalization capabilities across various underwater scenarios, as demonstrated by comprehensive experiments that show significant performance improvements over existing methods.
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2-11
No-Reference Image Quality Assessment Using Local Binary Patterns: A Comprehensive Performance Evaluation
One key in image quality assessment (IQA) is the design of image representations that can capture the changes in image structures caused by distortions. In the last decades, local binary patterns (LBP) have been proven to be a powerful tool as a statistical model for texture and local structure representation. LBP and its variants, as the texture descriptor with low computational complexity, have been applied widely and successfully in several specific applications, such as texture classification, face recognition and IQA. Generally, visible impairments alter the statistics of LBP descriptors, making it possible to measure degradation and then estimate image quality. However, only a few variants of LBP are applied in no-reference (NR) IQA and a few characteristics of LBP descriptors are explored for quality prediction, while some characteristics useful in the rest variants are ignored for IQA. Two previous studies give a review of 16/4 LBP operators and compare the performance with them separately in IQA application only on synthetic distorted data. To extend this work, we provide a review of LBP methodologies to assist the scientific community and new researchers, as well as explore more LBP descriptors in NR-IQA methods under various distortion conditions, particularly real-world cases. Specifically, we comprehensively review 30 widely-used or effective LBP-based operators, including recent variations. We then utilize a common framework for applying LBP descriptors in NR-IQA. The practicality of the reviewed descriptors is demonstrated and analyzed using experimental results under synthetic and authentic cases, indicating suitable LBP descriptors and characteristics for NR-IQA. Codes implementing the reviewed LBP operators is available at https://github.com/csxwan/Local-Binary-Patterns-for-IQA.
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 2-11
Open AccessNo-Reference Image Quality Assessment Using Local Binary Patterns: A Comprehensive Performance Evaluation
One key in image quality assessment (IQA) is the design of image representations that can capture the changes in image structures caused by distortions. In the last decades, local binary patterns (LBP) have been proven to be a powerful tool as a statistical model for texture and local structure representation. LBP and its variants, as the texture descriptor with low computational complexity, have been applied widely and successfully in several specific applications, such as texture classification, face recognition and IQA. Generally, visible impairments alter the statistics of LBP descriptors, making it possible to measure degradation and then estimate image quality. However, only a few variants of LBP are applied in no-reference (NR) IQA and a few characteristics of LBP descriptors are explored for quality prediction, while some characteristics useful in the rest variants are ignored for IQA. Two previous studies give a review of 16/4 LBP operators and compare the performance with them separately in IQA application only on synthetic distorted data. To extend this work, we provide a review of LBP methodologies to assist the scientific community and new researchers, as well as explore more LBP descriptors in NR-IQA methods under various distortion conditions, particularly real-world cases. Specifically, we comprehensively review 30 widely-used or effective LBP-based operators, including recent variations. We then utilize a common framework for applying LBP descriptors in NR-IQA. The practicality of the reviewed descriptors is demonstrated and analyzed using experimental results under synthetic and authentic cases, indicating suitable LBP descriptors and characteristics for NR-IQA. Codes implementing the reviewed LBP operators is available at https://github.com/csxwan/Local-Binary-Patterns-for-IQA.
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 31 (10709908) , 2495-2499
Spatial Adaptive Filter Network With Scale-Sharing Convolution for Image Demoiréing
Removing moiré patterns is a challenging task as it is a spatially varying degradation that varies in shape, color and scale. Existing image restoration models often rely on static convolutional neural networks (CNNs)-based architectures, and hence potentially suboptimal for addressing the diverse manifestations of moiré patterns across different images and spatial positions. To this end, we propose a spatially adaptive neural network for image demoiréing. This network introduces a dual-branch filter prediction module engineered to predict pixel-wise adaptive filters that can process moiré patterns of varying orientations and color-shift issues. To further tackle the challenge presented by scale variability, a scale-sharing convolution module is proposed, utilizing pixel-wise adaptive filters with multiple dilations to handle moiré patterns of different sizes but similar shapes effectively. Upon extensive evaluations of three benchmark datasets, our model consistently outperforms existing methods, yielding a PSNR improvement of over 0.37dB across all evaluated datasets and providing additional benefits in terms of model size.
- 9 September 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Signal Processing Letters
- Vol. 31 (10709908) , 2495-2499
Spatial Adaptive Filter Network With Scale-Sharing Convolution for Image Demoiréing
Removing moiré patterns is a challenging task as it is a spatially varying degradation that varies in shape, color and scale. Existing image restoration models often rely on static convolutional neural networks (CNNs)-based architectures, and hence potentially suboptimal for addressing the diverse manifestations of moiré patterns across different images and spatial positions. To this end, we propose a spatially adaptive neural network for image demoiréing. This network introduces a dual-branch filter prediction module engineered to predict pixel-wise adaptive filters that can process moiré patterns of varying orientations and color-shift issues. To further tackle the challenge presented by scale variability, a scale-sharing convolution module is proposed, utilizing pixel-wise adaptive filters with multiple dilations to handle moiré patterns of different sizes but similar shapes effectively. Upon extensive evaluations of three benchmark datasets, our model consistently outperforms existing methods, yielding a PSNR improvement of over 0.37dB across all evaluated datasets and providing additional benefits in terms of model size.
- 20 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 26 (15209210) , 7089-7100
CDINet: Content Distortion Interaction Network for Blind Image Quality Assessment
Perceptual image quality is related to content and distortion. Distortion classification is a common way to learn distortion information. How to extract distortion information consistent with human perception is a problem to be solved. Besides, the joint effect on image quality caused by the interplay of content and distortion has not been fully studied. In this paper, a novel Content Distortion Interaction Network (CDINet) is proposed for blind image quality assessment. Distortion representation are guided by content representation to learn quality-aware representation. CDINet consists of four components: a Distortion-Aware Module (DAM), a Content-Aware Module (CAM), an Asymmetric Content-Distortion Interaction (ACDI) module, and a quality regression module. The content representation and distortion representation are extracted respectively and fused interactively in CDINet. Specifically, with the assistance of image restoration, distortion representation consistent with human perception is learned. To further improve the ability in distortion representation, the DAM is used to construct the differences between the distorted image and its reference image. The proposed ACDI module enables the interaction of content and distortion representations to occur at different levels with less computational cost. Since the proposed CDINet considers the joint impact on image quality caused by the interplay of content and distortion, the predicted image qualities highly align with human perception. Comprehensive experiments on 8 benchmark datasets demonstrate that the proposed CDINet effectively extracts quality-aware representation, achieving state-of-the-art performance in evaluating both synthetically and authentically distorted images.
- 20 February 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Multimedia
- Vol. 26 (15209210) , 7089-7100
CDINet: Content Distortion Interaction Network for Blind Image Quality Assessment
Perceptual image quality is related to content and distortion. Distortion classification is a common way to learn distortion information. How to extract distortion information consistent with human perception is a problem to be solved. Besides, the joint effect on image quality caused by the interplay of content and distortion has not been fully studied. In this paper, a novel Content Distortion Interaction Network (CDINet) is proposed for blind image quality assessment. Distortion representation are guided by content representation to learn quality-aware representation. CDINet consists of four components: a Distortion-Aware Module (DAM), a Content-Aware Module (CAM), an Asymmetric Content-Distortion Interaction (ACDI) module, and a quality regression module. The content representation and distortion representation are extracted respectively and fused interactively in CDINet. Specifically, with the assistance of image restoration, distortion representation consistent with human perception is learned. To further improve the ability in distortion representation, the DAM is used to construct the differences between the distorted image and its reference image. The proposed ACDI module enables the interaction of content and distortion representations to occur at different levels with less computational cost. Since the proposed CDINet considers the joint impact on image quality caused by the interplay of content and distortion, the predicted image qualities highly align with human perception. Comprehensive experiments on 8 benchmark datasets demonstrate that the proposed CDINet effectively extracts quality-aware representation, achieving state-of-the-art performance in evaluating both synthetically and authentically distorted images.