In the development of radio frequency (RF) coils for better quality of the mini-type permanent magnetic resonance imager for using in the small animal imaging, the solenoid RF coil has a special advantage for permanent magnetic system based on analyses of various types of RF coils. However, it is not satisfied for imaging if the RF coils are directly used. By theoretical analyses of the magnetic field properties produced from the solenoid coil, the research direction was determined by careful studies to raise further the uniformity of the magnetic field coil, receiving coil sensitivity for signals and signal-to-noise ratio (SNR). The method had certain advantages and avoided some shortcomings of the other different coil types, such as, birdcage coil, saddle shaped coil and phased array coil by using the alloy materials (from our own patent). The RF coils were designed, developed and made for keeled applicable to permanent magnet-type magnetic resonance imager, multi-coil combination-type, single-channel overall RF receiving coil, and applied for a patent. Mounted on three instruments (25 mm aperture, with main magnetic field strength of 0.5 T or 1.5 T, and 50 mm aperture, with main magnetic field strength of 0.48 T), we performed experiments with mice, rats, and nude mice bearing tumors. The experimental results indicated that the RF receiving coil was fully applicable to the permanent magnet-type imaging system.
More and more medical devices can capture different features of human body and form three dimensional (3D) images. In clinical applications, usually it is required to fuse multiple source images containing different and crucial information into one for the purpose of assisting medical treatment. However, traditional image fusion methods are normally designed for two dimensional (2D) images and will lead to loss of the third dimensional information if directly applied to 3D data. Therefore, a novel 3D magnetic image fusion method was proposed based on the combination of newly invented beyond wavelet transform, called 3D band limited shearlet transformand (BLST), and four groups of traditional fusion rules. The proposed method was then compared with the 2D and 3D wavelet and dual-tree complex wavelet transform fusion methods through 4 groups of human brain T2* and quantitative susceptibility mapping (QSM) images. The experiments indicated that the performance of the method based on 3D transform was generally superior to the existing methods based on 2D transform. Taking advantage of direction representation, shearlet transform could effectively improve the performance of conventional fusion method based on 3D transform. It is well concluded, therefore, that the proposed method is the best among the methods based on 2D and 3D transforms.
Attention-deficit/hyperactivity disorder (ADHD) is one of the most common neuro-developmental disorders occurring in childhood, characterized by symptoms of age-inappropriate inattention, hyperactivity/impulsivity, and the prevalence is higher in boys. Although gray matter volume deficits have been frequently reported for ADHD children via structural magnetic resonance imaging, few of them had specifically focused on male patients. The present study aimed to explore the alterations of gray matter volumes in medicated-naive boys with ADHD via a relatively new voxel-based morphometry technique. According to the criteria of DSM-IV-TR, 43 medicated-naive ADHD boys and 44 age-matched healthy boys were recruited. The magnetic resonance image (MRI) scan was performed via a 3T MRI system with three-dimensional (3D) spoiled gradient recalled echo (SPGR) sequence. Voxel-based morphometry with diffeomorphic anatomical registration through exponentiated lie algebra in SPM8 was used to preprocess the 3D T1-weighted images. To identify gray matter volume differences between the ADHD and the controls, voxel-based analysis of whole brain gray matter volumes between two groups were done via two sample t-test in SPM8 with age as covariate, threshold at P＜0.001. Finally, compared to the controls, significantly reduced gray matter volumes were identified in the right orbitofrontal cortex (peak coordinates [-2,52,-25], t=4.01), and bilateral hippocampus (Left: peak coordinates [14,0,-18], t=3.61; Right: peak coordinates [-14,15,-28], t=3.64) of ADHD boys. Our results demonstrated obvious reduction of whole brain gray matter volumes in right orbitofrontal cortex and bilateral hippocampus in boys with ADHD. This suggests that the abnormalities of prefrontal-hippocam-pus circuit may be the underlying cause of the cognitive dysfunction and abnormal behavioral inhibition in medicated-naive boys with ADHD.
An adaptive inertia weight particle swarm algorithm is proposed in this study to solve the local optimal problem with the method of traditional particle swarm optimization in the process of estimating magnetic resonance (MR) image bias field. An indicator measuring the degree of premature convergence was designed for the defect of traditional particle swarm optimization algorithm. The inertia weight was adjusted adaptively based on this indicator to ensure particle swarm to be optimized globally and to avoid it from falling into local optimum. The Legendre polynomial was used to fit bias field, the polynomial parameters were optimized globally, and finally the bias field was estimated and corrected. Compared to those with the improved entropy minimum algorithm, the entropy of corrected image was smaller and the estimated bias field was more accurate in this study. Then the corrected image was segmented and the segmentation accuracy obtained in this research was 10% higher than that with improved entropy minimum algorithm. This algorithm can be applied to the correction of MR image bias field.
Attention deficit/hyperactivity disorder (ADHD) is a behavioral disorder syndrome found mainly in school-age population. At present, the diagnosis of ADHD mainly depends on the subjective methods, leading to the high rate of misdiagnosis and missed-diagnosis. To solve these problems, we proposed an algorithm for classifying ADHD objectively based on convolutional neural network. At first, preprocessing steps, including skull stripping, Gaussian kernel smoothing, et al., were applied to brain magnetic resonance imaging (MRI). Then, coarse segmentation was used for selecting the right caudate nucleus, left precuneus, and left superior frontal gyrus region. Finally, a 3 level convolutional neural network was used for classification. Experimental results showed that the proposed algorithm was capable of classifying ADHD and normal groups effectively, the classification accuracies obtained by the right caudate nucleus and the left precuneus brain regions were greater than the highest classification accuracy (62.52%) in the ADHD-200 competition, and among 3 brain regions in ADHD and the normal groups, the classification accuracy from the right caudate nucleus was the highest. It is well concluded that the method for classification of ADHD and normal groups proposed in this paper utilizing the coarse segmentation and deep learning is a useful method for the purpose. The classification accuracy of the proposed method is high, and the calculation is simple. And the method is able to extract the unobvious image features better, and can overcome the shortcomings of traditional methods of MRI brain area segmentation, which are time-consuming and highly complicate. The method provides an objective diagnosis approach for ADHD.
Heart diseases seriously threaten people's health. More and more functional evaluation of cardiac right ventricle has been considered in the clinical diagnosis in addition to the classical functional evaluation of cardiac left ventricle. It is very important to evaluate the functional parameters of right ventricle in clinical heart disease diagnosis, especially when the ejection fraction of left ventricle is very low. Right ventricular segmentation is needed for the functional evaluation. However, right ventricular segmentation has been difficult due to its thin myocardium, complex structure and significant individual variability. Cine cardiac magnetic resonance image is a golden standard in clinical functional evaluation of cardiac ventricle. In the present paper, we summarize the classic segmentation approaches, evaluation methods and their development, which can help the researchers in the related field have a quick and basic understanding to the right ventricle segmentation.
This paper performs a comprehensive study on the computer-aided detection for the medical diagnosis with deep learning. Based on the region convolution neural network and the prior knowledge of target, this algorithm uses the region proposal network, the region of interest pooling strategy, introduces the multi-task loss function: classification loss, bounding box localization loss and object rotation loss, and optimizes it by end-to-end. For medical image it locates the target automatically, and provides the localization result for the next stage task of segmentation. For the detection of left ventricular in echocardiography, proposed additional landmarks such as mitral annulus, endocardial pad and apical position, were used to estimate the left ventricular posture effectively. In order to verify the robustness and effectiveness of the algorithm, the experimental data of ultrasonic and nuclear magnetic resonance images are selected. Experimental results show that the algorithm is fast, accurate and effective.
The medical magnetic resonance (MR) image reconstruction is one of the key technologies in the field of magnetic resonance imaging (MRI). The compressed sensing (CS) theory indicates that the image can be reconstructed accurately from highly undersampled measurements by using the sparsity of the MR image. However, how to improve the image reconstruction quality by employing more sparse priors of the image becomes a crucial issue for MRI. In this paper, an adaptive image reconstruction model fusing the double dictionary learning is proposed by exploiting sparse priors of the MR image in the image domain and transform domain. The double sparse model which combines synthesis sparse model with sparse transform model is applied to the CS MR image reconstruction according to the complementarity of synthesis sparse and sparse transform model. Making full use of the two sparse priors of the image under the synthesis dictionary and transform dictionary learning, the proposed model is tackled in stages by the iterative alternating minimization algorithm. The solution procedure needs to utilize the synthesis and transform K-singular value decomposition (K-SVD) algorithms. Compared with the existing MRI models, the experimental results show that the proposed model can more efficiently improve the quality of the image reconstruction, and has faster convergence speed and better robustness to noise.
A multi-label based level set model for multiple sclerosis lesion segmentation is proposed based on the shape, position and other information of lesions from magnetic resonance image. First, fuzzy c-means model is applied to extract the initial lesion region. Second, an intensity prior information term and a label fusion term are constructed using intensity information of the initial lesion region, the above two terms are integrated into a region-based level set model. The final lesion segmentation is achieved by evolving the level set contour. The experimental results show that the proposed method can accurately and robustly extract brain lesions from magnetic resonance images. The proposed method helps to reduce the work of radiologists significantly, which is useful in clinical application.
In order to solve the pathological grading of hepatocellular carcinomas (HCC) which depends on biopsy or surgical pathology invasively, a quantitative analysis method based on radiomics signature was proposed for pathological grading of HCC in non-contrast magnetic resonance imaging (MRI) images. The MRI images were integrated to predict clinical outcomes using 328 radiomics features, quantifying tumour image intensity, shape and text, which are extracted from lesion by manual segmentation. Least absolute shrinkage and selection operator (LASSO) were used to select the most-predictive radiomics features for the pathological grading. A radiomics signature, a clinical model, and a combined model were built. The association between the radiomics signature and HCC grading was explored. This quantitative analysis method was validated in 170 consecutive patients (training dataset: n = 125; validation dataset, n = 45), and cross-validation with receiver operating characteristic (ROC) analysis was performed and the area under the ROC curve (AUC) was employed as the prediction metric. Through the proposed method, AUC was 0.909 in training dataset and 0.800 in validation dataset, respectively. Overall, the prediction performances by radiomics features showed statistically significant correlations with pathological grading. The results showed that radiomics signature was developed to be a significant predictor for HCC pathological grading, which may serve as a noninvasive complementary tool for clinical doctors in determining the prognosis and therapeutic strategy for HCC.