Yeqiu Li

Denoising by using multineural networks for medical X-ray imaging applications

In this paper, a new type of multineural network filter (MNNF) is presented that is trained for restoration and enhancement of the digital radiological images. In medical radiographices, noise has been categorized as quantum mottle, which is related to the incident X-ray exposure and artificial noise, which is caused by the grid, etc. MNNF consists of several neural network filters (NNFs). A novel analysis method is proposed to make the characteristics of the trained MNNF clearly. In the proposed method, a characteristics judgement system is presented to decide which NNF will be executed through the standard deviation value of pixels in the input region. The new approach was tested on nine clinical medical X-ray images and five synthesized noisy X-ray images. In all cases, the proposed MNNF produced better results in terms of peak signal-to-noise ratio (PSNR), mean-to-standard-deviation ratio (MSR) and contrast to noise ratio (CNR) measures than the original NNF, linear inverse filter and nonlinear median filter.

NOISE REMOVAL FOR MEDICAL X-RAY IMAGES IN MULTIWAVELET DOMAIN

When a signal is embedded in an additive Gaussian noise, its estimation is often done by finding a wavelet basis that concentrates the signal energy in few coefficients and then thresholding the noisy coefficients. However, in many practical problems such as medical X-ray image, astronomical and low-light images, the recorded data is not modeled by Gaussian noise but as the realization of a Poisson process. Multiwavelet is a new development to the body of wavelet theory. Multiwavelet simultaneously offers orthogonality, symmetry and short support which are not possible in scalar 2-channel wavelet systems. After reviewing this recently developed theory, a new theory and algorithm for denoising medical X-ray images using multiwavelet multiple resolution analysis (MRA) are presented and investigated in this paper. The proposed covariance shrink (CS) method is used to threshold wavelet coefficients. The form of thresholds is carefully formulated which is the key to more excellent results obtained in the extensive numerical simulations of medical image denoising compared to conventional methods

Removing Noise from Medical CR Image Using Multineural Network Filter Based on Noise Intensity Distribution

In this paper, a new type of multineural networks filter (MNNF) is presented that is trained for restoration and enhancement of the medical CR images. In medical CR image, noise has been categorized as quantum mottle, which is related to the incident X-ray exposure and artificial noise, which is caused by the grid, etc. MNNF consists of several neural network filters (NNFs). A novel analysis method is proposed to make the characteristics of the trained MNNF clearly. In the proposed method, a characteristics judgement system is presented to decide which NNF will be executed through the estimation of noise intensity calculated by maximum penalized likelihood estimator (MPLE). The new approach was tested on clinical medical X-ray image, synthesized noisy X-ray image and natural image. In all cases, the proposed MNNF produced better results in terms of mean square error (MSE) measure than MPLE, NNF and conventional wavelet BayesShrink (BS) methods.

A method of image denoising in the complex wavelet domain

In this paper, a new shrink theory and denoising algorithm for image with Gaussian noise based on complex wavelet transform is presented and investigated. We calculate threshold value by a moving window, we can obtain different threshold values for different coefficients using our method. We modify the noisy wavelet coefficients using bivariate shrinkage method, the shrinkage functions do not assume the independence of decompositional coefficients. In this paper, we propose the use of near-optimal thresholds and more suitable image denoising by using extensive numerical simulations.