• Journal of Information Processing Systems
• /
• v.16 no.5
• /
• pp.1129-1144
• /
• 2020

The thangka image inpainting method based on wavelet transform is not ideal for contour curves when the high frequency information is repaired. In order to solve the problem, a new image inpainting algorithm is proposed based on edge structural constraints and wavelet transform coefficients. Firstly, a damaged thangka image is decomposed into low frequency subgraphs and high frequency subgraphs with different resolutions using wavelet transform. Then, the improved fast marching method is used to repair the low frequency subgraphs which represent structural information of the image. At the same time, for the high frequency subgraphs which represent textural information of the image, the extracted and repaired edge contour information is used to constrain structure inpainting in the proposed algorithm. Finally, the texture part is repaired using texture synthesis based on the wavelet coefficient characteristic of each subgraph. In this paper, the improved method is compared with the existing three methods. It is found that the improved method is superior to them in inpainting accuracy, especially in the case of contour curve. The experimental results show that the hierarchical method combined with structural constraints has a good effect on the edge damage of thangka images.

• Journal of information and communication convergence engineering
• /
• v.4 no.4
• /
• pp.174-179
• /
• 2006

In this paper we present a text detection method inspired by wavelet packet analysis and improved fuzzy clustering algorithm(IAFC).This approach assumes that the text and non-text regions are considered as two different texture regions. The text detection is achieved by using wavelet packet analysis as a feature analysis. The wavelet packet analysis is a method of wavelet decomposition that offers a richer range of possibilities for document image. From these multi scale features, we adapt the improved fuzzy clustering algorithm based on the unsupervised learning rule. The results show that our text detection method is effective for document images scanned from newspapers and journals.

• Journal of the Korean Institute of Telematics and Electronics S
• /
• v.34S no.3
• /
• pp.55-62
• /
• 1997

For the effective removal of artifacts and the extraction of an improved evoked potential response, we propose the averaging method usin gthe shrinkag eof wavelet coefficients. The wavelet analysis decomposes the measured evoked potentials into scale coefficients with low frequency components and wavelet coefficients with high ones as a resolution level, respectively. and in the course of synthesis evoked potentials, the presented method shrinks the wavelet coefficients, and then reproduces the evoked potentials, and lastly averages it. We measured visual evoked potentials to simulate the averaging method using the shrinkage of wavelet coefficients, and compared it with aveaged signal. As a result of simulations, the proposed method gets improved VEP about 0.2-1.6dB in comparison with the averaging method with daubechies wavelet in the resolution level four.

• Nuclear Engineering and Technology
• /
• v.52 no.8
• /
• pp.1771-1776
• /
• 2020

Adverse effects in the measured gamma spectrum caused by radioactive statistical fluctuations, gamma ray scattering, and electronic noise can be reduced by energy spectrum denoising. Wavelet threshold denoising can be used to perform multi-scale and multi-resolution analysis on noisy signals with small root mean square errors and high signal-to-noise ratios. However, in traditional wavelet threshold denoising methods, there are signal oscillations in hard threshold denoising and constant deviations in soft threshold denoising. An improved wavelet threshold calculation method and threshold processing function are proposed in this paper. The improved threshold calculation method takes into account the influence of the number of wavelet decomposition layers and reduces the deviation caused by the inaccuracy of the threshold. The improved threshold processing function can be continuously guided, which solves the discontinuity of the traditional hard threshold function, avoids the constant deviation caused by the traditional soft threshold method. The examples show that the proposed method can accurately denoise and preserves the characteristic signals well in the gamma energy spectrum.

• IEMEK Journal of Embedded Systems and Applications
• /
• v.13 no.6
• /
• pp.313-319
• /
• 2018

We propose a post-processing algorithm to improve the quality of the RGB image generated by deep learning based colorization from the gray-scale image of an infrared camera. Wavelet fusion is used to generate a new luminance component of the RGB image luminance component from the deep learning model and the luminance component of the infrared camera. PSNR is increased for all experimental images by applying the proposed algorithm to RGB images generated by two deep learning models of SegNet and DCGAN. For the SegNet model, the average PSNR is improved by 1.3906dB at level 1 of the Haar wavelet method. For the DCGAN model, PSNR is improved 0.0759dB on the average at level 5 of the Daubechies wavelet method. It is also confirmed that the edge components are emphasized by the post-processing and the visibility is improved.

• Transactions of the Korean Society for Noise and Vibration Engineering
• /
• v.16 no.11 s.116
• /
• pp.1149-1157
• /
• 2006

Donoho et al. suggested a wavelet thresholding denoising method based on discrete wavelet transform. This paper proposes an improved denoising method using a new thresholding function based on translation-invariant wavelet for underwater acoustic measurement. The conventional wavelet thresholding denoising method causes Pseudo-Gibbs phenomena near singularities due to the lack of translation-invariant of the wavelet basis. To suppress Pseudo-Gibbs phenomena, a denoising method combining a new thresholding function based on the translation-invariant wavelet transform is proposed in this paper. The new thresholding function is a modified hard-thresholding to each node according to the discriminated threshold so as to reject unknown external noise and white gaussian noise. The experimental results show that the proposed method can effectively eliminate noise, extract characteristic information of radiated noise signals.

• The Journal of the Acoustical Society of Korea
• /
• v.20 no.2E
• /
• pp.25-29
• /
• 2001

In this paper, we propose a new method for selecting the threshold on wavelet packet denoising. In selecting threshold, the method using median is not efficient. Because this method can not recover unvoiced signal corrupted by noise. So we partition a speech signal corrupted by noise into the pure noise section and voiced section using autocorrelation and entropy. The autocorrelation and entropy can reflect disorder of noise. The new method yields more improved denoising effect. Especially unvoiced signal is very nicely reconstructed, and SNR is improved.

• Proceedings of the IEEK Conference
• /
• summer
• /
• pp.244-248
• /
• 2004

This paper proposes a new noise cancellation method for speech recognition in noise environments. We determine the time adaptive threshold value using standard deviations of wavelet coefficients after wavelet transform by frames. The time adaptive threshold value is set up by using sum of standard deviations of wavelet coefficients in cA3 and weighted cD1. cA3 coefficients represent the voiced sound with lower frequency components and cD1 coefficients represent the unvoiced sound with higher frequency components. In experiments, we removed noise after adding white Gaussian noise and colored noise to original speech. The proposed method improved SNR and MSE more than wavelet transform and wavelet packet transform does. As a result of speech recognition experiment using noise speech DB, recognition performance is improved by $2\sim4\;\%.$

• Proceedings of the KOSOMBE Conference
• /
• v.1996 no.11
• /
• pp.229-232
• /
• 1996

we propose the shrinkage of wavelet coefficients and the averaging method. The wavelet analysis decomposes the measured evoked potentials into scale coefficients and wavelet coefficients as a resolution level, respectively. And in the course of synthesis of evoked potentials, the presented method shrinks the wavelet coefficients, and then reproduces the evoked potentials and lastly averages it. we measured VEP signal to simulate the presented method, and compared it with averaged signal and LMS algorithm. As a result of simulations, the proposed method gets improved VEP about 0.2-1.6db in comparison with the result of averaging method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.13 no.5
• /
• pp.2434-2453
• /
• 2019

To improve the separation performance of time-frequency overlapped radar and communication signals from a single channel, this paper proposes an effective separation method based on an improved empirical wavelet transform (EWT) that introduces a fast boundary detection mechanism. The fast boundary detection mechanism can be regarded as a process of searching, difference optimization, and continuity detection of the important local minima in the Fourier spectrum that enables determination of the sub-band boundary and thus allows multiple signal components to be distinguished. An orthogonal empirical wavelet filter bank that was designed for signal adaptive reconstruction is then used to separate the input time-frequency overlapped signals. The experimental results show that if two source components are completely overlapped within the time domain and the spectrum overlap ratio is less than 60%, the average separation performance is improved by approximately 32.3% when compared with the classic EWT; the proposed method also improves the suitability for multiple frequency shift keying (MFSK) and reduces the algorithm complexity.