[Design of high performance DSP-based gradient calculation module for MRI].

A gradient calculation module based on high performance DSP was designed to meet the needs of digital MRI spectrometer. According to the requirements of users, this apparatus can achieve rotation transformation, pre-emphasis, shimming and other gradient calculation functions in a single chip of DSP. It then outputs gradient waveform data of channel X, Y, Z and shimming data of channel B0. Experiments show that the design has good versatility and can satisfy the functional, speed and accuracy requirements of MRI gradient calculation. It provides a practical gradient calculation solution for the development of digital spectrometer.

A new approach to accelerate DR image enhancement based on CUDA / 中国医疗器械杂志

Multiscale pyramid image enhancement algorithm is an usual way to enhance the Digital Radiography (DR) images. However, the process of enhancement takes much of time because of the fine resolution of DR images. A new method of accelerating DR image enhancement based on Compute Unified Device Architecture (CUDA) is presented in this paper. This method completes a large amount of convolution operations in spatial domain involved in the multiscale pyramid image enhancement algorithm by using the Graphic Processing Unit (GPU). The experimental results show that the proposed method is very efficient for accelerating DR image enhancement.

[DR image denoising based on Laplace-Impact mixture model].

A novel DR image denoising algorithm based on Laplace-Impact mixture model in dual-tree complex wavelet domain is proposed in this paper. It uses local variance to build probability density function of Laplace-Impact model fitted to the distribution of high-frequency subband coefficients well. Within Laplace-Impact framework, this paper describes a novel method for image denoising based on designing minimum mean squared error (MMSE) estimators, which relies on strong correlation between amplitudes of nearby coefficients. The experimental results show that the algorithm proposed in this paper outperforms several state-of-art denoising methods such as Bayes least squared Gaussian scale mixture and Laplace prior.

[The design method of gradient coils in permanent MRI].

According to the features of the open permanent MRI system, this paper designs the gradient magnetic field, in which the designed parameters are optimized. The simulation experiments demonstrate the effectiveness of the algorithm.

[An adaptive scaling hybrid algorithm for reduction of CT artifacts caused by metal objects].

A new adaptively hybrid filtering algorithm is proposed to reduce the artifacts caused by metal in CT image. Firstly, the method is used to preprocess the projection data of metal region and is reconstruct by filtered back projection (FBP) method. Then the expectation maximization algorithm (EM) is performed on the iterative original metal project data. Finally, a compensating procedure is applied to the reconstructed metal region. The simulation result has demonstrated that the proposed algorithm can remove the metal artifacts and keep the structure information of metal object effectively. It ensures that the tissues around the metal will not be distorted. The method is also computational efficient and effective for the CT images which contains several metal objects.

[Algorithm of digital demodulation and filtering in MRI spectrometers].

In the design of RF receiver of the digital MRI spectrometer console, a digital demodulation and filtering algorithm is presented in this thesis. The MR signals are firstly converted to digital signals by the A/D converter, and then quadrature-demodulated from high frequency carrier wave. The CIC filter, half-band filter and linear phase FIR filter are designed to process the cascaded filtering and decimation of the demodulated signals. This method achieves a satisfying processing speed and filtering effect, and also reduces the data size obviously. The experiment based on the permanent magnetism resonance imaging system proves its effectiveness and practicability.

[An algorithm of nodule detection based on high resolution CT images].

Combining the speed of two-dimensional detection with the precision of three-dimensional detection, an automatic algorithm based on high resolution CT images is proposed to identify nodules in this paper. Nodule candidates are extracted by a two-dimensional convergence index (CI) filter, then a three-dimensional Hessian matrix detection filter is introduced to reduce false positive lung nodules, Experiments show that the algorithm is effective with a sensitivity of 90% and the false positive lung nodules per slice is 0.33.

[Design of a digital spectrometer for the MRI system].

A spectrometer is one of the most important parts in a Magnetic Resonance Imaging (MRI) system. This paper describes the design of a digital MRI spectrometer. It is constructed on a PXI platform with several data acquisition boards and a high-resolution timing board. All functions of a MRI spectrometer are realized by the specially- designed software. The software architecture and its implementing details are discussed and experimental results are introduced.

[Finite element method-based optimal design of MRI permanent magnet].

This paper presents a simulation method to study and improve the technology of designing magnets. With the finite element method, it analyzes the magnetic field distribution of the magnet model constructed by CAD software. Based on the distribution characteristics of magnetic field, the redundancy parts of the magnet configuration are removed accurately. The experiment results show that this method can significantly lighten the magnet.

[Design of the image browser for PACS image workstation].

The design of PACS image workstation based on DICOM3.0 is introduced in the paper, then the designing method of the PACS image browser based on the control system theory is presented,focusing on two main units:DICOM analyzer and the information mapping transformer.

[Research on designing MRI permanent magnet].

Based on the theory of magnet circuits, the paper introduces the process of designing MRI (Magnetic Resonance Imaging) permanent magnet. The measurement of magnet properties is improved by simulation technology. In order to increase the magnetic field homogeneity. We have designed a shim-loop and a trapeziform shim-board to optimize the magnetic field distribution. The results of both simulation analysis and experiments show that the innovative design improves magnetic field properties significantly and the magnet structure accords with the technology requirements.

[Controlling of the high-voltage generator in a direct radiography system].

A proposal is presented to improve the original complicated control of the exposure parameters in the direct radiography system (DR). The interface circuitry of the high-voltage generator is redesigned using RS-232 serial COM port. The data and signals are transmitted between the personal computer and the high-voltage generator in the serial mode. A program package is designed to realize the control of the exposure parameters of the high-voltage generator in DR on the PC machine. Experiment results have shown that the proposed console operates steadily and it is capable of providing a more convenient operation console for the direct radiography system.

[Wavelet threshold denoising algorithm for CR images].

A wavelet threshold denoising algorithm for CR images is put forward here . A noised CR image is decomposed into wavelet coefficients, which are processed by the algorithm, and the denoised CR image is reconstructed based on the processed coefficients. Examples are too presented to demonstrate the efficiency of the algorithm on denoising and, maintaining the detail of the CR images.

A novel digital magnetic resonance imaging spectrometer.

Spectrometer is the essential part of magnetic resonance imaging (MRI) system. It controls the transmitting and receiving of signals. Many commercial spectrometers are now available. However, they are usually costly and complex. In this paper, a new digital spectrometer based on PCI extensions for instrumentation (PXI) architecture is presented. Radio frequency (RF) pulse is generated with the method of digital synthesis and its frequency and phase are continuously tunable. MR signal acquired by receiver coils is processed by digital quadrature detection and filtered to get the k-space data, which avoid the spectral distortion due to amplitude and phase errors between two channels of traditional detection. Compared to the conventional design, the presented spectrometer is built with general PXI platform and boards. This design works in a digital manner with features of low cost, high performance and accuracy. The experiments demonstrate its efficiency.

Classifying ECoG/EEG-based motor imagery tasks.

The multichannel electrocorticogram (ECoG)/electroencephalogram (EEG) signals are commonly used to classify two kinds of motor imagery (MI) tasks. In this paper, the ECoG and EEG data sets are composed of training and test data, which are recorded during different time/days. Power spectral density (PSD) is selected as features; Fisher discriminant analysis (FDA) and common spatial patterns (CSP) are used to filter redundancy; K-Nearest-Neighbor (KNN) classifier is applied to classify MI tasks; and a new function R (k) is presented to estimate the value of k. Using these methods, we obtain the predictive accuracy of MI tasks based on ECoG data (which is 92%) and EEG data (which is 81%). The results show that we can effectively classify two kinds of MI tasks based on EEG as well as ECoG.

[An automatic color correction algorithm for digital human body sections].

OBJECTIVE: To find a new approach to improve the uniformity of color parameters for images data of the serial sections of the human body. METHOD: An auto-color correction algorithm in the RGB color space based on a standard CMYK color chart was proposed. The gray part of the color chart was auto-segmented from every original image, and fifteen gray values were attained. The transformation function between the measured gray value and the standard gray value of the color chart and the lookup table were obtained. In RGB color space, the colors of images were corrected according to the lookup table. RESULT: The color of original Chinese Digital Human Girl No. 1 (CDH-G1) database was corrected by using the algorithm with Matlab 6.5, and it took 13.475 s to deal with one picture on a personal computer. CONCLUSION: Using the algorithm, the color of the original database is corrected automatically and quickly. The uniformity of color parameters for corrected dataset is improved.

[Fast extraction of iso-surface in surface rendering of medical images].

In this paper a new direct volume rendering method is presented for fast extraction of iso-surface by adopting the idea from the Shear-Warp algorithm. By creating the sorted volumetric data from the original volume data and specifying a value range of data which determines the part of the sorted volumetric data traversed, the amount of volume data traversed would be reduced obviously and the extraction operation of iso-surface would be very fast. In addition, we can adjust the value range to obtain the different rendering speed and image quality according to the purpose in application. Moreover, the proposed algorithm will not output any intermediate data after the sorted volumetric data being produced. Therefore, it is possible to realize the rapid 3D surface reconstruction for medical images on the personal computer without the support of any hardware accelerator.

A method for artifact correction due to demodulation phase errors in magnetic resonance imaging.

In the process that MR signal is quadrature demodulated, the initial phases of sinusoid and cosinusoid used as demodulation reference signal exist errors, which results in artifacts in the image. It is typically overcome through precise control of the initial phases with the help of the special hardware, but the errors are not reduced completely. In this work, a method based on reference scan is proposed. Reference scan without frequency encoding gradient and phase encoding gradient is executed before FSE sequence, and an echo train is acquired. The phase errors of sinusoid and cosinusoid that are used to demodulate each echo in the reference echo train are calculated. Then the k-space data of image is corrected by these errors and artifacts are removed. The experiments using 0.35T MRI system demonstrate the effectiveness of this method.

[A rapid method of eddy current compensation in magnetic resonance imaging systems].

Imaging objects are spatially encoded by gradient magnetic fields in magnetic resonance imaging systems. The eddy current caused by rapid switches of gradient fields will result in artifacts in the images. A method of eddy current compensation based on pre-emphasis of gradient current is presented in this thesis. The compensation parameters are acquired rapidly utilizing Faraday's induction theorem and data fitting method. The experiments prove that the method is efficient for reduction of the debugging time and for the improvement of the image quality.

Volume rendering based on sorted volumetric data.

OBJECTIVE: To describe a spatial data structure called sorted volumetric data structure which speeds up the volume rendering and will not affect image quality. METHOD: Before volume rendering we transfer each slice of volume into an array indexed by the value of voxels, element in which stores the position of each voxel. According to the opacity transfer function, we can affirm the value ranges of voxels that are non-transparent. Therefore we translate, resample and composite only the voxels in the value ranges. By marching through the sorted arrays, we locate the non-transparent voxels rapidly and skip all voxels that are transparent. The sorted data structures need not recompute whenever the opacity transfer function changes and are not restricted on the opacity transfer function. RESULT: The method presented in this paper has been implemented in a standard personal computer. The rendering time for CT head volume data is less than 1 s that is satisfied with the diagnostic purpose. CONCLUSION: The concept of proposed algorithm is simple, easy to realize and it is not recomputed whenever the opacity transfer function changes. By using the sorted volumetric data structure, we speed up the volume rendering without affecting image quality.