The on-going evolution of technology of capturing and generation video, sound, RF, etc. signals results in very HDR data to be stored and processed. The analysis of human perception of different signals is showing tha...
详细信息
To address the problem of low accuracy and poor stability of bearing diagnostic models under strong background noise, a bearing fault image recognition method is proposed that reduces the randomness of the model by av...
详细信息
Coronary Atherosclerosis is recognized as the predominant cause of Chronic Heart Failure. The standard diagnostic modality for Coronary Atherosclerosis involves echocardiography (ECO);however, the accessibility of suc...
详细信息
The paper describes a new method of noisy image restoration based on amplification of low-frequency components in high-frequency subbands, or details of fast wavelet transform (FWT). It makes a difference with respect...
详细信息
ISBN:
(数字)9798331518752
ISBN:
(纸本)9798331518769
The paper describes a new method of noisy image restoration based on amplification of low-frequency components in high-frequency subbands, or details of fast wavelet transform (FWT). It makes a difference with respect to well-known wavelet filtering methods which are typically based on soft thresholding of wavelet coefficients belonging to high-frequency subbands of a noisy image. At the same time, the low-frequency component, or approximation can be used to extract the low-frequency components corresponding the details. It allows modifying the details of the FWT by adding the extracted low-frequency components that leads to increasing the signal-to-noise ratios in subbands. The paper contains the experimental results of modeling proving the effectiveness of the suggested method comparing to some widely used filtering methods, too. The results of modeling have shown that the proposed method allows improving quality of restored image by increasing both the signal-to-noise ratio and significantly reducing the number of ringing artifacts.
Unmanned underwater vehicles generally rely on high quality visual data and low latency for the applications of monitoring, exploration and search. Several methods have been proposed to improve underwater visibility b...
详细信息
Currently, image generation and synthesis have remarkably progressed with generative models. Despite photo-realistic results, intrinsic discrepancies are still observed in the frequency domain. The spectral discrepanc...
详细信息
ISBN:
(纸本)1577358872
Currently, image generation and synthesis have remarkably progressed with generative models. Despite photo-realistic results, intrinsic discrepancies are still observed in the frequency domain. The spectral discrepancy appeared not only in generative adversarial networks but in diffusion models. In this study, we propose a framework to effectively mitigate the disparity in frequency domain of the generated images to improve generative performance of both GAN and diffusion models. This is realized by spectrum translation for the refinement of image generation (STIG) based on contrastive learning. We adopt theoretical logic of frequency components in various generative networks. The key idea, here, is to refine the spectrum of the generated imagevia the concept of image-to-image translation and contrastive learning in terms of digital signalprocessing. We evaluate our framework across eight fake image datasets and various cutting-edge models to demonstrate the effectiveness of STIG. Our framework outperforms other cutting-edges showing significant decreases in FID and log frequency distance of spectrum. We further emphasize that STIG improves image quality by decreasing the spectral anomaly. Additionally, validation results present that the frequency-based deepfake detector confuses more in the case where fake spectrums are manipulated by STIG.
Discrete wavelet Transform (DWT) is an ubiquitous mathematical technique used in a wide range of applications. The fundamental advantages of DWT are its high compression ratio, lack of blocking artifacts, and strong t...
详细信息
ISBN:
(纸本)9798350346787
Discrete wavelet Transform (DWT) is an ubiquitous mathematical technique used in a wide range of applications. The fundamental advantages of DWT are its high compression ratio, lack of blocking artifacts, and strong time & frequency domain localization. This proposed work is an efficient implementation of (9,7) lifting based 1D, 2D, and 3D DWTs, where the hardware software codesign with partial reconfiguration (PR) has made a significant reduction in resource utilization. Additionally, the suggested approach yields a very low mean square error (MSE), resulting in a substantially higher peak signal to noise ratio (PSNR). The constant multiply-add unit based proposed work with single core implementation achieves 80% decrease in the number of look up tables (LUT) and 32% decrease in number of slices utilized when compared with the variable multiply-add unit based implementation. To prove the architecture's originality, it is tested for six distinct applications such as mono audio, stereo audio, grayscale image, colour image, gray scale video, and colour video. The implementation of the work is done on Zynq 7000 (XC7Z020CLG484-1) field programmable gate array (FPGA) with Xilinx vivado.
Satellite image denoising is of utmost importance in various applications weather monitoring, flood control and crop monitoring focusing on enhancing the visual quality of affected images. Denoising is a contemporary ...
详细信息
The wavelet transform has emerged as a powerful tool in deciphering structural information within images. And now, the latest research suggests that combining the prowess of wavelet transform with neural networks can ...
详细信息
ISBN:
(纸本)1577358872
The wavelet transform has emerged as a powerful tool in deciphering structural information within images. And now, the latest research suggests that combining the prowess of wavelet transform with neural networks can lead to unparalleled image deraining results. By harnessing the strengths of both the spatial domain and frequency space, this innovative approach is poised to revolutionize the field of imageprocessing. The fascinating challenge of developing a comprehensive framework that takes into account the intrinsic frequency property and the correlation between rain residue and background is yet to be fully explored. In this work, we propose to investigate the potential relationships among rainfree and residue components at the frequency domain, forming a frequency mutual revision network (FMRNet) for image deraining. Specifically, we explore the mutual representation of rain residue and background components at frequency domain, so as to better separate the rain layer from clean background while preserving structural textures of the degraded images. Meanwhile, the rain distribution prediction from the low-frequency coefficient, which can be seen as the degradation prior is used to refine the separation of rain residue and background components. Inversely, the updated rain residue is used to benefit the low-frequency rain distribution prediction, forming the multi-layer mutual learning. Extensive experiments demonstrate that our proposed FMRNet delivers significant performance gains for seven datasets on image deraining task, surpassing the state-of-the-art method ELFormer by 1.14 dB in PSNR on the Rain100L dataset, while with similar computation cost. Code and retrained models are available at https://***/kuijiang94/FMRNet.
Norm calculation for vectors and matrices is a fundamental mathematical operation that has been widely utilized across various domains, including signalprocessing and wireless communication systems. In this paper, a ...
详细信息
ISBN:
(数字)9798350377866
ISBN:
(纸本)9798350377873;9798350377866
Norm calculation for vectors and matrices is a fundamental mathematical operation that has been widely utilized across various domains, including signalprocessing and wireless communication systems. In this paper, a pioneering approach is introduced for implementing norm calculations pertaining to vectors and matrices. The key idea involves the application of the in memory computing (IMC) technique, where computations are executed through a series of memristive devices. More precisely, the fundamental multiplication and addition operations leverage the inherent properties of memristor devices, following the principles of Ohm's law and Kirchhoff's current law. This innovative approach not only enhances flexibility by leveraging the unique capabilities of IMC in programming memristors but also emphasizes processing efficiency. Consequently, through the reduction of computational complexity and latency, it outperforms traditional implementation approaches.
暂无评论