- Advanced Image and Video Retrieval Techniques
- Image Retrieval and Classification Techniques
- Advanced Steganography and Watermarking Techniques
- Visual Attention and Saliency Detection
- Video Analysis and Summarization
- Image and Video Quality Assessment
- Advanced Vision and Imaging
- Distributed systems and fault tolerance
- Advanced Image Fusion Techniques
- Video Coding and Compression Technologies
- Image Enhancement Techniques
- Full-Duplex Wireless Communications
- Digital Media Forensic Detection
- Software System Performance and Reliability
- Distributed and Parallel Computing Systems
- Domain Adaptation and Few-Shot Learning
- Cooperative Communication and Network Coding
- Scientific Computing and Data Management
- Generative Adversarial Networks and Image Synthesis
- Network Security and Intrusion Detection
- Video Surveillance and Tracking Methods
- Energy Harvesting in Wireless Networks
Guangxi Normal University
2019-2024
Huazhong University of Science and Technology
2024
Virginia Commonwealth University
2010
Lehigh University
2007
Pennsylvania State University
2004
Image quality assessment (IQA) is an important task of image processing and has diverse applications, such as super-resolution reconstruction, transmission monitoring systems. This paper proposes a perceptual hashing algorithm with complementary color wavelet transform (CCWT) compressed sensing (CS) for reduced-reference (RR) IQA. The CCWT exploited to decompose input into different sub-bands. Since the calculation uses all channels without discarding any information, distortions introduced...
Image hashing is a useful technique of many multimedia systems, such as image authentication, copy detection, tampering detection and quality assessment (IQA). However, most schemes do not make desirable performance IQA. To tackle this, new scheme with deep texture features proposed for reduced-reference (RR) In the hashing, are calculated from discrete cosine transform coefficients three-order tensor stacked by feature maps generated pre-trained ResNet18. Texture extracted Gray-level...
Copy detection is a key task of image copyright protection. Most robust hashing schemes do not make satisfied performance copy yet. To address this, scheme with deep features and Meixner moments proposed for detection. In the hashing, global are extracted by applying tensor Singular Value Decomposition (t-SVD) to three-order constructed in DWT domain feature maps calculated pre-trained VGG16. Since slightly disturbed digital operations, stable thus desirable robustness guaranteed. Moreover,...
Abstract Image hashing is an efficient technique of many multimedia systems, such as image retrieval, authentication and copy detection. Classification between robustness discrimination one the most important performances hashing. In this paper, we propose a robust with singular values quaternion value decomposition (QSVD). The key contribution innovative use QSVD, which can extract stable discriminative features from CIE L*a*b* color space. addition, block are viewed point in Cartesian...
Image hashing is an efficient technique of multimedia processing for many applications, such as image copy detection, authentication, and social event detection. In this study, the authors propose a novel with visual attention model invariant moments. An important contribution weighted DWT (discrete wavelet transform) representation by incorporating called Itti saliency into LL sub‐band. Since can efficiently extract map reflecting regions focus, perceptual robustness proposed achieved....
Robust video hashing is an effective method of copy detection. But most existing schemes do not make classifications; thus their detection performances are unsatisfactory. To address these problems, we propose a robust based on deep feature and the quaternion generic Fourier descriptor (QGFD) for In proposed scheme, entropy-weighted secondary frame calculated by weighting all frames in group via color entropy. Because entropy can reflect information, incorporation into each guarantees...
Most user cooperation protocols work in a timesharing manner, where each transmits its own message and relays for the other at different segments of time slot. We develop new scheme to send these messages simultaneously using network coding. show that coding is more tolerant poor inter-user channels than time-sharing, achieves better overall performance. generalize multi-user, multi-slot framework. Under this framework, we reaps diversity order provides effective channel time-sharing schemes.
Robust hashing is a powerful technique for processing large-scale images. Currently, many reported image schemes do not perform well in balancing the performances of discrimination and robustness, thus they cannot efficiently detect copies, especially copies with multiple distortions. To address this, we exploit global local invariant features to develop novel robust copy detection. A critical contribution feature calculation by gray level co-occurrence moment learned from saliency map...
Robust hashing is a useful technique for the image applications of watermarking, authentication, quality assessment and copy detection. This paper proposes new robust detection by using local tangent space alignment (LTSA). A key contribution weighted visual map computation based on difference Gaussian (DOG) attention model. The can provide proposed method with good robustness. Another feature learning via LTSA from matrix in discrete cosine transform domain. As it maintain geometric...
Workflow systems are popular in daily business processing. Since vulnerability cannot be totally removed from a workflow management system, successful attacks always happen and may inject malicious tasks or incorrect data into the system. Referring to further corrupt more objects which comprises integrity level of This problem efficiently solved by existing defense mechanisms, such as access control, intrusion detection, checkpoints. In this paper, we propose practical solution for online...
Robustness is an important property of image hashing. Most the existing hashing algorithms do not reach good robustness against large-angle rotation. Aiming at this problem, we jointly exploit visual attention model and ring partition to design a novel hashing, which can make rotation robustness. In proposed called PFT (Phase spectrum Fourier Transform) used detect saliency map preprocessed image. The LL sub-band then divided into concentric circles invariant by partition, means variances...
Abstract Image hashing is an effective technology for extensive image applications, such as retrieval, authentication and copy detection. This paper designs a new scheme based on saliency map sparse model. The major contributions are twofold. first contribution the construction of weighted representation by combining visual attention model called Itti matrix color vector angle (CVA). Since can efficiently detect CVA fully captures information image, they contribute to visually robust...
Human capacity for "lifelong learning" encompasses a continuous process of acquiring knowledge, adapting to new environments, and developing skills throughout one's life. In order bridge the gap between human intelligence artificial intelligence, an increasing number researchers have begun explore concept lifelong learning within field machine learning, also referred as or incremental learning. Incremental enables learn from stream data, thereby achieving capability. can be categorized into...
Dividing a Screen Content Image (SCI) with complex components into pictorial and textual regions for predicting scores is one of the common Quality Assessment (SCIQA) methods. However, how to efficiently leverage features predict quality no-reference SCIQA still needs be explored. In addition, statistical analysis reveals that labels SCIs present distribution. Therefore, both distribution need considered in SCIQA. This paper proposes method unifying features. One contribution proposed...
Copy detection is a key task of video copyright protection. This paper presents robust hashing with non-negative tensor factorization (NTF) for copy detection. In the presented scheme, secondary frames are computed from preprocessed by assigning weights to all within group based on color entropy. Next, fed into pre-trained MobileNetV2 and then NTF exploited compress three-order constructed stacking output feature maps hash construction. Experiments conducted publicly available datasets...
Video hashing is an efficient technique for tasks like copy detection and retrieval. This paper utilizes canonical polyadic (CP) decomposition Hahn moments to design a robust video hashing. The first significant contribution the secondary frame construction. It uses three weighted techniques generate frames each group, which can effectively capture features of from different aspects thus improves discrimination. Another deep feature extraction via ResNet50 CP decomposition. use provide rich...
Abstract This paper proposes a novel video hashing with tensor robust Principal Component Analysis (PCA) and Histogram of Optical Flow (HOF) for copy detection. In the proposed hashing, is divided into some groups. For each group, low-rank secondary frame constructed from component decomposed by applying PCA to group. Since can well indicate spatial-temporal intrinsic structure group it slightly disturbed digital operations, feature extraction frames discriminative stable. Next, spatial...
Recovery from attacks has been extensively studied at the database transaction level and application in recent years. To recover compromised transactions, compensating redoing transactions need to be conducted under concurrency control restrictions. Under a multi-tier service architecture, level, attack recovery more restrictions introduced by either dependencies among activities or specifications. Thus, architecture introduces challenges problem. In this study, authors describe problems...