Article
Mathematics, Applied
Alban Gossard, Frederic de Gournay, Pierre Weiss
Summary: A recent trend in the signal/image processing literature is the optimization of Fourier sampling schemes for specific datasets. This paper explains why choosing optimal non Cartesian Fourier sampling patterns is a difficult nonconvex problem by addressing two optimization issues: the existence of a combinatorial number of spurious minimizers and the vanishing gradient effect for high frequencies. The paper concludes by demonstrating how the use of large datasets and stochastic gradient algorithms with a variable metric can alleviate these issues.
Article
Optics
Muqian Wen, John Houlihan
Summary: Resampling by interpolation is a traditional method for processing interferograms from non-uniformly sampled Fourier transform spectrometers, while the non-uniform fast Fourier transform (NUFFT) is an overlooked alternative approach. Through experiments on a high-resolution interferometer and various optical sources, these two methods are compared. It is found that NUFFT is comparable to interpolation in spectral profile shape and spectral noise levels, but superior in spectral amplitude and computer performance. Moreover, a novel implementation of NUFFT is presented and analyzed.
OPTICS COMMUNICATIONS
(2023)
Article
Engineering, Electrical & Electronic
Deyun Wei, Jun Yang
Summary: A novel estimation algorithm is proposed for non-uniform discrete Fourier transform (NUDFT) in this paper, which achieves compression of sparse signals and accurate detection of frequencies.
IEEE TRANSACTIONS ON SIGNAL PROCESSING
(2022)
Article
Computer Science, Information Systems
Cheng Guo, Wanping Liu, Ximeng Liu, Yinghui Zhang
Summary: This article presents a method called SEND for secure similarity search over encrypted non-uniform and high-dimensional datasets. By combining SSE with locality-sensitive hashing and utilizing selective hashing and query set distribution hiding techniques, SEND achieves high search quality in terms of recall and precision, and is proven secure against adaptively chosen query attacks in the standard model.
IEEE TRANSACTIONS ON CLOUD COMPUTING
(2022)
Article
Multidisciplinary Sciences
Yao Jia, Chen Cui, Ahmed A. Abd El-Latif
Summary: This paper proposes a new image hashing method based on histogram reconstruction to enhance the sensitivity of histograms to changes in pixel position and ensure robustness against common content preservation attacks. The algorithm leverages rotational symmetry and invariance to rotation operations, enabling resistance to arbitrary angles of rotation. Experimental results demonstrate the method's strong performance in robustness and discrimination compared to established algorithms. Additionally, a receiver operating characteristic curve analysis further confirms the superior overall performance of the proposed image hash method.
Article
Multidisciplinary Sciences
Tieyu Zhao, Yingying Chi
Summary: The study introduced the multiple-parameter fractional Fourier transform (MPFRFT) as an image encryption algorithm, but found many of its keys to be invalid, reducing security. By reformulating the MPFRFT and analyzing using different basis functions, it was discovered that the effective keys were extremely limited, highlighting the security risk of key invalidation in extended encryption methods based on MPFRFT.
Article
Mathematics, Interdisciplinary Applications
Hui Zhao, Bing-Zhao Li
Summary: This brief report examines the problem of unlimited sampling of high dynamic non-bandlimited signals in the Fourier domain based on the fractional Fourier transform. It proposes a mathematical signal model for unlimited sampling and uses the annihilation filtering method to estimate arbitrary folding time. A novel fractional Fourier domain unlimited sampling theorem is obtained.
FRACTAL AND FRACTIONAL
(2023)
Article
Computer Science, Artificial Intelligence
Yao Xiao, Wei Zhang, Xiangguang Dai, Xiangqin Dai, Nian Zhang
Summary: In this paper, a robust supervised hashing framework, called Robust Supervised Discrete Hashing (RSDH), is proposed based on the Cauchy loss function and Supervised Discrete Hashing (SDH) to learn a subspace consisted of binary codes. RSDH can reduce outliers and noise of the hashing codes and achieve better retrieval performance.
Review
Computer Science, Information Systems
Abdul Subhani Shaik, Ram Kumar Karsh, Mohiul Islam, Rabul Hussain Laskar
Summary: This paper discusses the importance of image hashing technology in multimedia security applications, focusing on key challenges and design parameters in hashing schemes. It summarizes the existing literature on hashing-based image authentication techniques, explores different algorithm performance and comparisons between datasets.
MULTIMEDIA TOOLS AND APPLICATIONS
(2022)
Article
Mathematics, Interdisciplinary Applications
Sri Sulasteri, Mawardi Bahri, Nasrullah Bachtiar, Jeffry Kusuma, Agustinus Ribal
Summary: The main objective of this study is to find the solution of the generalized heat and generalized Laplace equations using the fractional Fourier transform, which is a general form of the solution of these equations using the classical Fourier transform. The solution is also formulated using a sampling formula related to the fractional Fourier transform. The study introduces the fractional Fourier transform, collects related theorems and essential properties, and derives several results related to the sampling formula. Several examples are presented to demonstrate the effectiveness and powerfulness of the proposed method compared to the classical Fourier transform method.
FRACTAL AND FRACTIONAL
(2023)
Article
Automation & Control Systems
An Wang, Mobarakol Islam, Mengya Xu, Hongliang Ren
Summary: In this study, a curriculum-based Fourier domain adaptation (Curri-AFDA) method is proposed for medical image segmentation, aiming to enhance the autonomy of computer-aided diagnosis and intervention systems. By introducing curriculum learning strategy and training-time chained augmentation mixing, the method achieves good adaptation and generalization performance in multiple testing domains.
IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING
(2023)
Article
Computer Science, Information Systems
S. Qasim Abbas, Fawad Ahmed, Yi-Ping Phoebe Chen
Summary: The proposed PIH scheme in this paper combines DCT and NRLBP to compute image hash for improved robustness against malicious distortions and the ability to detect localized tampering as small as 3% of the original image size. This scheme outperforms existing hashing schemes and exhibits resilience against non-malicious distortions.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Engineering, Electrical & Electronic
Deyun Wei, Jun Yang
Summary: The discrete fractional Fourier transform is a useful tool in non-stationary signal processing. This paper proposes a two-dimensional sparse fractional Fourier transform (2D SFRFT) algorithm, which achieves the lowest runtime and sample complexity compared to existing methods. The algorithm is improved to be robust by analyzing errors due to noises, and its effectiveness is demonstrated in various applications.
Article
Computer Science, Theory & Methods
Xinran Li, Chuan Qin, Zichi Wang, Zhenxing Qian, Xinpeng Zhang
Summary: In this paper, a unified performance evaluation method for perceptual image hashing schemes is proposed, which contains six modules and assigns scores to each module using order relationship analysis (ORA). Experimental results demonstrate that this method is practical and effective for evaluating the performance of perceptual image hashing schemes.
IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY
(2022)
Article
Computer Science, Hardware & Architecture
Yajuan Sun, Gautam Srivastava
Summary: This paper proposes a digital watermarking algorithm based on locality-sensitive hashing to improve the robustness of attacking defence for video images. The experimental results demonstrate that the proposed method has good invisibility and high accuracy in identifying and countering image tampering and pseudo-authentication attacks.
MOBILE NETWORKS & APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Chuan Qin, Jinchuan Hu, Fengyong Li, Zhenxing Qian, Xinpeng Zhang
Summary: JPEG image encryption converts the original JPEG image into a noise-like image without any useful information. Existing schemes may not balance file size increment and encryption security. To address this, we propose a novel scheme that predicts DC coefficients and encrypts the histogram of prediction errors, ensuring a small file size increment. We also implement permutation for RS pairs in each DCT block to distort image contents. Extensive experiments demonstrate our scheme's compatibility, small file size increment, and superior security performance compared to existing schemes.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Computer Science, Information Systems
Fengyong Li, Hengjie Zhu, Chuan Qin
Summary: This paper proposes a reversible data hiding scheme in encrypted images by combining median prediction and bit plan cycling-XOR, achieving a better balance among good visual quality, large embedding capacity and high security performance.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Engineering, Electrical & Electronic
Mingji Yu, Heng Yao, Chuan Qin, Xinpeng Zhang
Summary: This paper proposes a novel comprehensive measure to evaluate the performance of stream-cipher-based reversible data hiding algorithms, taking into consideration the influence of encryption schemes. By categorizing the characteristics of encryption and embedding processes, and using correlation, redundancy, embedding rate, computation complexity, visual quality, and algorithm stability as evaluation indexes, this method provides a more reasonable comparison of RDHEI algorithms. The experimental results demonstrate the guiding significance of the proposed method for current RDHEI algorithm selection.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Computer Science, Information Systems
Fengyong Li, Zhenjia Pei, Weimin Wei, Jing Li, Chuan Qin
Summary: In this study, we propose a method for image forgery detection by fusing multi-resolution features and designing a tamper-guided dual self-attention module. Experimental results demonstrate that compared to existing schemes, our method can achieve pixel-level forgery localization and image-level forgery detection simultaneously, while maintaining higher detection accuracy and stronger robustness.
SECURITY AND COMMUNICATION NETWORKS
(2022)
Article
Engineering, Electrical & Electronic
Fengyong Li, Zhen Qi, Xinpeng Zhang, Chuan Qin
Summary: This article introduces a new JPEG reversible data hiding scheme that improves image quality and reduces file size expansion by optimizing distortion models and reassigning frequency priorities.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Mingji Yu, Heng Yao, Chuan Qin, Xinpeng Zhang
Summary: Reversible data hiding (RDH) has been studied for over two decades and can be divided into RDH and reversible data hiding in encrypted images (RDHEI) depending on the application scenario. Most studies have focused on gray-scale images, with relatively few conducted on color images and even fewer on palette images. Therefore, a new framework specifically for palette images is proposed, which includes a route selection algorithm followed by a color table reconstruction. Experimental results demonstrate the improvement brought by this route selection to existing RDH/RDHEI methods in palette images.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Yalan Qin, Chuan Qin, Xinpeng Zhang, Donglian Qi, Guorui Feng
Summary: In this paper, the authors investigate three challenging problems in the field of incomplete multi-view representation learning and propose a novel framework called NIM-Nets to address these challenges. NIM-Nets can fully utilize incomplete data from different views to produce a consistent, informative, and robust multi-view shared representation. The authors also introduce definitions of robustness and completeness for incomplete multi-view representation learning.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2023)
Article
Computer Science, Artificial Intelligence
Gejian Zhao, Chuan Qin, Heng Yao, Yanfang Han
Summary: This study proposes a novel protection scheme for deep neural network (DNN) models, which utilizes a self-embedding fragile watermark to ensure model integrity and parameter recovery. The experimental results demonstrate that the proposed scheme achieves satisfactory tampering detection and parameter recovery with low device requirements, and it can be effectively applied to various existing DNN models.
PATTERN RECOGNITION LETTERS
(2022)
Article
Computer Science, Hardware & Architecture
Guopeng Gao, Chuan Qin, Yaodong Fang, Yuanding Zhou
Summary: In this work, a new perceptual authentication hashing model for digital images based on contrastive unsupervised learning is proposed. By utilizing a contrastive augmentation structure and an integrated loss function, the proposed scheme achieves robustness to unknown manipulations and outperforms other existing schemes in performance.
Article
Engineering, Electrical & Electronic
Chunqiang Yu, Xianquan Zhang, Chuan Qin, Zhenjun Tang
Summary: This study proposes a new reversible data hiding method using Chinese remainder theorem-based secret sharing and hybrid coding, which achieves high embedding capacity and good security properties.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Computer Science, Information Systems
Gejian Zhao, Fengyong Li, Heng Yao, Chuan Qin
Summary: This paper proposes a robust fingerprinting scheme for video content authentication. It extracts key spatio-temporal features from the input video and maps them to the corresponding fingerprint using a two-dimensional attention mechanism and spatio-temporal weighted fusion. Experimental results show that the proposed scheme achieves superior performances in terms of robustness and discrimination.
JOURNAL OF VISUAL COMMUNICATION AND IMAGE REPRESENTATION
(2023)
Article
Computer Science, Artificial Intelligence
Jian Li, Xiaobo Zhang, Bin Ma, Chuan Qin, Chunpeng Wang
Summary: In this paper, a method of protecting a device from being identified through image capture is proposed. This method involves suppressing the device-related fingerprint, PRNU, in an image by modifying it in the domain of integer wavelet transform. The suppressed PRNU is then embedded into the image using reversible data hiding for perfect recovery of the original image. Extensive experiments on MICHE-I and VISION datasets demonstrate that this method addresses the shortcomings of existing techniques and exhibits strong scalability.
EXPERT SYSTEMS WITH APPLICATIONS
(2023)
Article
Geochemistry & Geophysics
Jiayi Wu, Chuan Qin, Yanli Ren, Guorui Feng
Summary: This study presents a novel semantic segmentation network structure that integrates prototype information with global edge information to achieve more accurate prototype-matching results. In addition, a comprehensive weighted loss function is designed to monitor the training process and help overcome challenges. Results of performance comparison with existing few-shot semantic segmentation methods demonstrate the superiority of the proposed method.
IEEE GEOSCIENCE AND REMOTE SENSING LETTERS
(2023)
Article
Computer Science, Artificial Intelligence
Xinran Li, Mengqi Guo, Zichi Wang, Jian Li, Chuan Qin
Summary: This article proposes a powerful image hashing method for encrypted images, which can extract hash sequences from the encrypted version of an image without decryption, to protect the privacy of image content. The method encrypts the plaintext image using the Paillier cryptosystem, and then extracts the hash sequence through Walsh-Hadamard transform and histogram statistics. In the encrypted domain, image owners can efficiently authenticate their images through a cloud server without disclosing the image content. Compared with traditional hashing methods designed for plaintext domain, our method achieves satisfactory privacy-preserving performance and better image authentication performance.
IEEE TRANSACTIONS ON EMERGING TOPICS IN COMPUTATIONAL INTELLIGENCE
(2023)
Article
Computer Science, Hardware & Architecture
Chuan Qin, Shengyan Gao, Xinpeng Zhang, Guorui Feng
Summary: This research designs an effective image watermarking attack method based on the conditional generative adversarial nets (CGAN). It constructs a targeted and combined loss function according to the network structure of CGAN, which can ensure an acceptable visual quality of attacked image and make the watermark extraction fail simultaneously. Experimental results demonstrate the effectiveness of this method for some state-of-the-art deep robust image watermarking methods with transferability. As an attack method, it can serve as an evaluation standard to measure the robustness of watermarking.
Article
Engineering, Electrical & Electronic
Zou Deyue, Gao Siyu, Li Xinyue, Zhao Wanlong
Summary: In this paper, an integrated navigation/communication signal is proposed by combining Cyclic Code Shift Keying (CCSK) as a communication signal with traditional Global Navigation Satellite System (GNSS) signal. A fuzzy judging algorithm and segmenting correlation domain are used to improve the transmission robustness of the integrated signal. Additionally, a strategy is proposed to solve the incompatibility problem between the integrated signal and binary communication system using interleaving coding and dual-thresholds optimization. The proposed algorithm significantly improves the fault-tolerant rate and reduces the Bit Error Rate (BER), enhancing the reliability of the communication system.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Guoxing Huang, Yunfei Xiang, Shibiao Deng, Yu Zhang, Jingwen Wang
Summary: This paper proposes a dual-channel cooperative sub-Nyquist sampling system for LFM-BPSK hybrid modulated signal. The system can solve the frequency ambiguity problem caused by low sampling rate and estimate the phases of binary symbols and the positions of discontinuities through self-mixing technology.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
A. Yousefian Darani, Y. Khedmati Yengejeh, G. Navarro, H. Pakmanesh, J. Sharafi
Summary: The development and use of network technology have brought society into the internet world, where communication through digital data takes place. However, the protection of digital resources has become a significant challenge, which steganography is seen as a possible solution for. This paper presents an image steganography algorithm that hides a grayscale secret image within an RGB color cover image, with a genetic algorithm used to optimize the selection of suitable hiding places. The proposed method is empirically demonstrated to be effective in terms of transparency, security, and resistance.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Xiao Tan, Zhiwei Yang, Xianghai Li, Yongfei Mao, Di Jiang
Summary: A novel gridless SR-STAP algorithm applicable to a non-ULA with AAPEs is proposed in this study. By establishing the ANM-STAP model and iteratively updating the clutter subspace and AAPEs, the performance of STAP is improved.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Irfan Manisali, Okyanus Oral, Figen S. Oktem
Summary: In this paper, a novel non-iterative deep learning-based reconstruction method for real-time near-field MIMO imaging is proposed. The method achieves high image quality with low computational cost at compressive settings through two stages of processing. A large-scale dataset is also developed for training the neural networks. The effectiveness of the method is demonstrated through experimental data and extensive simulations.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Haibin Li, Dengchao Wu, Wenming Zhang, Cunjun Xiao
Summary: Workplace safety accidents are a pervasive issue worldwide, with 67.95% of construction accidents attributed to the lack of helmet-wearing. Existing helmet detection algorithms suffer from underperformance in real-world scenarios due to various challenges. This study introduces a lightweight helmet detection algorithm, YOLO-PL, which achieves state-of-the-art performance by optimizing network structure and incorporating lightweight modules.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Fen Liu, Jianfeng Chen, Kemeng Li, Jisheng Bai, Weijie Tan, Chang Cai, Muhammad Saad Ayub
Summary: In this paper, a semi-tensor product-based multi-modal factorized multilinear (STP-MFM) pooling method is proposed for information fusion in sentiment analysis. Experimental results demonstrate that the proposed method outperforms baselines in terms of accuracy, training speed, and model complexity.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Xiaoyu Zhao, Xiaoqian Chen, Zhiqiang Gong, Wen Yao, Yunyang Zhang, Xiaohu Zheng
Summary: In this paper, a contrastive enhancement approach is proposed to mine latent prototypes from background regions and leverage latent classes to improve the utilization of similarity information between prototype and query features. The proposed modules, including a latent prototype sampling module and a contrastive enhancement module, significantly improve the performance of state-of-the-art methods for few-shot segmentation.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Xiaoyong Lyu, Baojin Liu, Wenbing Fan, Zhi Quan
Summary: This paper provides an in-depth analysis of the channel estimate model (CEM) in passive radar using orthogonal frequency division multiplexing (OFDM) waveforms. The authors propose a new CEM that takes into consideration the influence of inter-carriers interference (ICI). The theoretical analysis and simulation results support the effectiveness of the new CEM and the proposed method for canceling ICI.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Ruisheng Rana, Jinping Wang, Bin Fang, Weiming Yang
Summary: This paper introduces an improved neighborhood preserving embedding method (NPEAE), which utilizes a linear autoencoder to achieve more accurate and effective data projection from high-dimensional space to low-dimensional space. NPEAE performs better in recognition accuracy compared to other comparative methods.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Yuehao Guo, Xianpeng Wang, Jinmei Shi, Lu Sun, Xiang Lan
Summary: This article presents a fast real-valued tensor propagator method for FDA-MIMO radar. The method improves estimation accuracy by utilizing the original structural information of multidimensional data and eliminates the high computational complexity of high-order singular value decomposition. The proposed algorithm achieves parameter estimation at low snapshots and has lower computational complexity than other algorithms at high snapshots.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Bang Huang, Wen-Qin Wang, Weijian Liu, Shunsheng Zhang, Yizhen Jia
Summary: This paper studies the robust moving target detection problem in FDA-MIMO radar with an unknown covariance matrix in Gaussian clutter. The proposed approach adopts the subspace method and proposes three robust adaptive detectors, which are validated to be effective.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Asutosh Kar
Summary: In this paper, a class of variable step size adaptive algorithms is developed for hybrid narrow-band active noise control (HNANC) systems and compared with the existing state-of-the-art methods. Two algorithms are proposed for HNANC systems operating in multiple noise environments, and their performance is analyzed. The results demonstrate significant improvement in noise reduction compared to the counterparts.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Shouqi Wang, Zhigang Feng
Summary: With the rapid development of artificial intelligence and sensor technology, intelligent fault diagnosis methods based on deep learning are widely used in industrial production. In practical applications, complex noise environments and huge model parameters affect the performance and cost-effectiveness of diagnostic models. In this paper, a lightweight intelligent fault diagnosis model using multi-sensor data fusion is proposed to address these issues. By processing vibration signals from different sensors of rolling bearings using variational mode decomposition (VMD) and designing unique grayscale feature maps based on each intrinsic modal function (IMF) component, the proposed model achieves high accuracy diagnosis in noisy environments while meeting the requirements of small, light, and fast production. The ultra-lightweight GoogLeNet model (ULGoogLeNet) is constructed by adjusting the traditional GoogLeNet structure, and the ultra-lightweight subspace attention module (ULSAM) is introduced to reduce model parameters and enhance feature extraction capability. Experimental results on two datasets demonstrate the effectiveness and superiority of the proposed method.
DIGITAL SIGNAL PROCESSING
(2024)
Article
Engineering, Electrical & Electronic
Jichen Yang, Fangfan Chen, Yu Cheng, Pei Lin
Summary: Recently, there has been significant attention on multi-speaker multimedia speaker recognition (MMSR). This study explores innovative techniques for integrating audio and visual cues from the front-end representations of both speaker's voice and face. Experimental results show that these techniques achieve considerable improvements in the performance of MMSR.
DIGITAL SIGNAL PROCESSING
(2024)
