Article
Computer Science, Artificial Intelligence
Subhadeep Koley, Hiranmoy Roy, Soumyadip Dhar, Debotosh Bhattacharjee
Summary: This paper proposes a new face image descriptor called LDCTBP for illumination- and modality-invariant face recognition. The proposed LDCTBP is generated by encoding illumination-normalized DCT coefficients into a binary pattern. Experiments on three datasets demonstrate the superiority of the proposed method over state-of-the-art approaches. Additionally, the integration of LDCTBP with a light-weight CNN highlights the importance of handcrafted features in CNN training.
PATTERN ANALYSIS AND APPLICATIONS
(2023)
Article
Chemistry, Multidisciplinary
Su-Yeon Lim, Min-Kyeong Choi, Yung-Lyul Lee
Summary: This paper proposes a method to enhance the fractional interpolation of reference samples in VVC intra prediction. The method utilizes additional interpolation filters that use integer-positioned reference samples based on their frequency information. Four alternative interpolation filters, including 8-tap/4-tap DCT-IFs and 4-tap/8-tap SIFs, are introduced along with an interpolation filter selection method based on the high-frequency ratio calculated from the one-dimensional transform of the reference samples. The proposed frequency-based Adaptive Filter achieves overall Bjontegaard Delta (BD) rate gains of -0.16%, -0.13%, and -0.09% for Y, Cb, and Cr components, respectively, compared with VVC.
APPLIED SCIENCES-BASEL
(2023)
Article
Computer Science, Information Systems
Anil Kumar Yadav, R. K. Pateriya, Nirmal Kumar Gupta, Punit Gupta, Dinesh Kumar Saini, Mohammad Alahmadi
Summary: Face recognition systems have improved human-computer interactions, with PCA-SVM showing better generalization capabilities and avoiding overfitting compared to PCA-ANN. However, PCA-SVM is ineffective in cases of poor lighting, long hair, or items covering the face. This study proposes a novel PCA-SVM-based model to overcome PCA-ANN's recognition issues and enhance face detection.
CMC-COMPUTERS MATERIALS & CONTINUA
(2022)
Article
Computer Science, Information Systems
Abdulhakeem O. Mohammed, Haval I. Hussein, Ramadhan J. Mstafa, Adnan M. Abdulazeez
Summary: With the rise of the Internet of Things and smart devices, it has become easier to access and distribute digital data. However, these advancements have also made it easier for intruders to violate copyright protection, commit identity theft, and compromise privacy. To address these issues, image watermarking, specifically a blind watermarking approach based on joint Discrete Cosine Transform (DCT) and Discrete Wavelet Transform (DWT) for RGB color images, is proposed in this paper. The approach demonstrates enhanced watermark invisibility and robustness, meeting the main requirements of image watermarking.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Computer Science, Software Engineering
Xianjun Han, Huabin Wang, Hongyu Yang, Xuejun Li
Summary: This paper proposes a small data-assisted face image illumination normalization method, which improves generalization ability by training the network model on a small number of image pairs and using different network parameters for different face images. Experiments show that the method performs well in normalization even when only 500 face image pairs are used for training.
COMPUTERS & GRAPHICS-UK
(2021)
Article
Computer Science, Artificial Intelligence
Lusia Rakhmawati, Wirawan Wirawan, Suwadi Suwadi, Claude Delpha, Pierre Duhamel
Summary: The paper proposes a novel reversible robust watermarking scheme based on the JPEG image, which minimizes total distortion for the watermarked image by adaptively selecting embedding strengths and coefficients and using an advanced block selection strategy to correspond to the distortion. Experimental results show that the method can maintain good visual quality while being robust to geometrical and image processing attacks.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Huan Liu, Xinglin Zhang, Haobin Dong, Zheng Liu, Xiangyun Hu
Summary: The article introduces an MAD method based on energy-concentrated discrete cosine wavelet transform (WT), which improves the signal-to-noise ratio (SNR) of the magnetic anomaly by extracting the principal components of the scale signal in wavelet transformation. The experimental results demonstrate that the proposed WT-DCT achieves the highest SNR improvement and the best structural similarity (SSIM) improvement among four commonly used methods.
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT
(2023)
Article
Computer Science, Information Systems
Ekta Soni, Arpita Nagpal, Puneet Garg, Placido Rogerio Pinheiro
Summary: The large number of COVID-19 patients in hospitals has made it difficult to screen heart patients. Telecardiology combined with CNN algorithm and compression technology can accurately classify arrhythmias while reducing data storage space, although at the cost of increased computational time.
Article
Computer Science, Hardware & Architecture
Yuheng Jiang, Jiajia Chen
Summary: This article proposes a new algorithm to derive Int-DCT with good coding performance and low hardware cost by addressing the tradeoff between coding performance and implementation cost. Through multiobjective optimization and efficient search, the algorithm reduces the implementation area and power cost while causing negligible impact to coding performance in comparison to other competing algorithms in ASIC experiments.
IEEE TRANSACTIONS ON COMPUTER-AIDED DESIGN OF INTEGRATED CIRCUITS AND SYSTEMS
(2022)
Article
Computer Science, Information Systems
Mannan Hassan, Muhammad Suhail Shaikh, Munsif Ali Jatoi
Summary: This study addresses the challenge of illumination invariance in facial recognition systems by employing normalization of lighting and histogram-based image enhancement methods. By utilizing logarithmic transformations and DCT, the performance is significantly improved in face images with high illumination dissimilarities.
MULTIMEDIA SYSTEMS
(2022)
Article
Computer Science, Information Systems
Xiaohui Kuang, Xianfeng Gao, Lianfang Wang, Gang Zhao, Lishan Ke, Quanxin Zhang
Summary: Deep learning models are widely used in various fields, but vulnerable to security threats. This study presents a novel targeted attack method against state-of-the-art object detection models, showing significant effects.
INFORMATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Achyut Shankar
Summary: In this paper, an Artificial Intelligence based Effective Data Interpretation Approach (AI-EDIA) is proposed to improve underwater wireless sensor network communication and reduce computational time in IoT platform. The AI-EDIA utilizes discrete cosine transform (DCT) with frequency modulation multiplexing (FMM) for underwater acoustic communication. Experimental results show that AI-EDIA decreases energy usage and delay rate to 0.45 s.
ARTIFICIAL INTELLIGENCE REVIEW
(2023)
Article
Engineering, Electrical & Electronic
Elham Esmaeili, Farshad Pesaran, Nabiollah Shiri
Summary: Sophisticated systems utilize approximate circuits for energy management. A gate-level structure is introduced to create a novel imprecise full adder (FA) cell using the gate diffusion input (GDI) strategy. The FA is utilized in three 4:2 compressors for partial product reduction in a new 8-bit inexact multiplier.
INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS
(2023)
Article
Computer Science, Information Systems
Omer Faruk Saritas, Serkan Ozturk
Summary: In this article, a novel hybrid Contourlet Transform (CT) and Discrete Cosine Transform (DCT) based blind and robust color image watermarking method is presented to improve the robustness and invisibility of watermarks. Experimental results demonstrate the effectiveness of the proposed method in protecting the transparency of watermarked images against attacks.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Chemistry, Multidisciplinary
Chun-Chieh Lin, Pei-Chann Chang, Ping-Heng Tsai
Summary: This study proposes a dual-adaptive approach based on DCT for the removal of baseline wander in ECG records. The method accurately calculates the cardiac fundamental frequency and uses DCT coefficients to distinguish between baseline wander and ECG. A new filtering technique based on DCT is then applied to suppress baseline wander. Experimental results demonstrate significant improvements in performance when using this dual-adaptive approach.
APPLIED SCIENCES-BASEL
(2022)
Article
Computer Science, Information Systems
Bhawna Ahuja, Virendra P. Vishwakarma
Summary: A novel multikernel deterministic extreme learning machine (ELM) and its variants are developed in this paper for classification of non-linear problems. By analytically calculating parameters and utilizing data fusion proficiency of multiple kernel learning, a deterministic solution with high face recognition success rate is achieved.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Article
Multidisciplinary Sciences
Sahil Dalal, Virendra P. Vishwakarma
Summary: This study utilizes ECG waveforms as a biometric identification measure, conducting experiments with databases such as MLII, UCI repository arrhythmia, and PTBDB, using features extracted through multi-cumulants and optimized parameters with genetic algorithm for classification using kernel extreme learning machine. The approach achieved excellent classification rates on different databases, showcasing its efficacy compared to existing methods.
SCIENTIFIC REPORTS
(2021)
Article
Information Science & Library Science
Sahil Dalal, Virendra P. Vishwakarma
Summary: This paper presents a new algorithm for face image recognition using extreme learning machine, genetic algorithm, and PCA, achieving improvement in recognition accuracy across various face databases compared to standard methods.
JOURNAL OF INFORMATION & OPTIMIZATION SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Bhawna Ahuja, Virendra P. Vishwakarma
Summary: This paper presents a deterministic extreme learning machine for neural network with feedforward architecture formulated with multiple kernel learning, and further enhances the approach by incorporating Gray level co-occurrence matrix (GLCM) for multi-modal feature extraction.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Mathematics, Applied
Bhawna Ahuja, Virendra P. Vishwakarma
Summary: This paper introduces kernel ELM, evaluates the optimal solution using genetic algorithm, and improves performance, with experimental results showing it to be more accurate and efficient than existing algorithms in the literature.
JOURNAL OF DISCRETE MATHEMATICAL SCIENCES & CRYPTOGRAPHY
(2021)
Article
Computer Science, Information Systems
Varsha Sisaudia, Virendra P. Vishwakarma
Summary: This paper introduces a novel approach for copyright protection of multi-spectral images using PSO and KELM techniques, showing promising results compared to other state-of-art approaches.
MULTIMEDIA TOOLS AND APPLICATIONS
(2021)
Proceedings Paper
Computer Science, Information Systems
Virendra P. Vishwakarma, Sahil Dalal
ADVANCES IN VLSI, COMMUNICATION, AND SIGNAL PROCESSING
(2020)
Proceedings Paper
Computer Science, Information Systems
Virendra P. Vishwakarma, Sudesh Yadav
ADVANCES IN VLSI, COMMUNICATION, AND SIGNAL PROCESSING
(2020)
Proceedings Paper
Computer Science, Artificial Intelligence
Divya Verma, Virendra Prasad Vishwakarma, Sahil Dalal
ADVANCES IN BIOINFORMATICS, MULTIMEDIA, AND ELECTRONICS CIRCUITS AND SIGNALS
(2020)
Proceedings Paper
Computer Science, Artificial Intelligence
Sahil Dalal, Virendra P. Vishwakarma, Sanchit Kumar
INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE AND DATA SCIENCE
(2020)
Proceedings Paper
Computer Science, Artificial Intelligence
Sahil Dalal, Virendra P. Vishwakarma
INTERNATIONAL CONFERENCE ON COMPUTATIONAL INTELLIGENCE AND DATA SCIENCE
(2020)
Article
Mathematics, Applied
Gargi Mishra, Virendra P. Vishwakarma
JOURNAL OF DISCRETE MATHEMATICAL SCIENCES & CRYPTOGRAPHY
(2020)