A soft-segmentation visualization scheme for magnetic resonance images

Prevalent visualization tools exploit gray value distribution in images through modified histogram equalization and matching technique, referred to as the window width/window level-based method, to improve visibility and enhance diagnostic value. The window width/window level tool is extensively used in magnetic resonance (MR) images to highlight tissue boundaries during image interpretation. However, the identification of different regions and distinct boundaries between them based on gray-level distribution and displayed intensity levels is extremely difficult because of the large dynamic range of tissue intensities inherent in MR images. We propose a soft-segmentation visualization scheme to generate pixel partitions from the histogram of MR image data using a connectionist approach and then generate selective visual depictions of pixel partitions using pseudo color based on an appropriate fuzzy membership function. By applying the display scheme in clinical examples in this study, we could demonstrate additional overlapping regions between distinct tissue types in healthy and diseased areas (in the brain) that could help improve the tissue characterization ability of MR images.     

Influence of age and osteoporosis on calcaneus trabecular bone structure: a preliminary in vivo MRI study by quantitative texture analysis

Recent developments in high-resolution MR imaging techniques have opened up new perspectives for structural characterization of trabecular bone by non-invasive methods. In this study, 3-D MR imaging was performed on 17 healthy volunteers and 6 osteoporotic patients. Two different MR sequences were used to evaluate the impact on MR acquisition on texture analysis results. Images were analyzed with four automated methods of texture analysis (grey level histogram, cooccurrence, runlength and gradient matrices) enabling quantitative analysis of grey level intensity and distribution within three different regions of interest (ROI). Texture analysis is not very frequently used since the interpretation of the large number of calculated parameters is difficult. We applied multiparametric data analyses such as principal component analysis (CFA) and hierarchical ascending classification (HAC) to determine the relevant parameters to differentiate between three sets of images (healthy young volunteers, healthy postmenopaused and osteoporotic patients). The results suggest that relevant texture information (depending on the ROI localization in the calcaneus) can be extracted from calcaneus MR images to evaluate osteoporosis and age effects on trabecular bone structure if strictly the same acquisition sequences are used for all patients' examination.     

Comparison of automated and visual texture analysis in MRI: characterization of normal and diseased skeletal muscle

Automated magnetic resonance imaging (MRI) texture analysis was compared with visual MRI analysis for the diagnosis of skeletal muscle dystrophy in 14 healthy and 17 diseased subjects. MRI texture analysis was performed on 8 muscle regions of interest (ROI) using four statistical methods (histogram, co-occurrence matrix, gradient matrix, runlength matrix) and one structural (mathematical morphology) method. Nine senior radiologists assessed full leg transverse slice images and proposed a diagnosis. The 59 extracted texture parameters for each ROI were statistically analyzed by Correspondence Factorial Analysis. Non-parametric tests were used to compare diagnoses based on automated texture analysis and visual analysis. Texture analysis methods discriminated between healthy volunteers and patients with a sensitivity of 70%, and a specificity of 86%. Comparison with visual analysis of MR images suggests that texture analysis can provide useful information contributing to the diagnosis of skeletal muscle disease.     

A novel, fast entropy-minimization algorithm for bias field correction in MR images

A novel, fast entropy-minimization algorithm for bias field correction in magnetic resonance (MR) images is suggested to correct the intensity inhomogeneity degradation of MR images that has become an increasing problem with the use of phased-array coils. Four important modifications were made to the conventional algorithm: (a) implementation of a modified two-step sampling strategy for stacked 2D image data sets, which included reducing the size of the measured image on each slice with a simple averaging method without changing the number of slices and then using a binary mask generated by a histogram threshold method to define the sampled voxels in the reduced image; (b) improvement of the efficiency of the correction function by using a Legendre polynomial as an orthogonal base function polynomial; (c) use of a nonparametric Parzen window estimator with a Gaussian kernel to calculate the probability density function and Shannon entropy directly from the image data; and (d) performing entropy minimization with a conjugate gradient method. Results showed that this algorithm could correct different types of MR images from different types of coils acquired at different field strengths very efficiently and with decreased computational load.     

Pediatric abdominal masses: diagnostic accuracy of diffusion weighted MRI

Purpose

To retrospectively identify apparent diffusion coefficient (ADC) values of pediatric abdominal mass lesions, to determine whether measured ADC of the lesions and signal intensity on diffusion-weighted (DW) images allow discrimination between benign and malignant mass lesions.

Materials and Methods

Approval for this retrospective study was obtained from the institutional review board. Children with abdominal mass lesions, who were examined by DW magnetic resonance imaging (MRI) were included in this study. DW MR images were obtained in the axial plane by using a non breath-hold single-shot spin-echo sequence on a 1.5-T MR scanner. ADCs were calculated for each lesion. ADC values were compared with Mann–Whitney U test. Receiver operating characteristic curve analysis was performed to determine cut-off values for ADC. The results of visual assessment on b800 images and ADC map images were compared with chi-square test.

Results

Thirty-one abdominal mass lesions (16 benign, 15 malignant) in 26 patients (15 girls, 11 boys, ranging from 2 days to 17 years with 6.9 years mean) underwent MRI. Benign lesions had significantly higher ADC values than malignant ones (P<.001). The mean ADCs of malignant lesions were 0.84±1.7×10⁻³ mm²/s, while the mean ADCs of the benign ones were 2.28±1.00×10⁻³ mm²/s. With respect to cutoff values of ADC: 1.11×10⁻³ mm²/s, sensitivity and negative predictive values were 100%, specificity was 78.6% and positive predictive value was 83.3%. For b800 and ADC map images, there were statistically significant differences on visual assessment. All malignant lesions had variable degrees of high signal intensity whereas eight of the 16 benign ones had low signal intensities on b800 images (P<.001). On ADC map images, all malignant lesions were hypointense and most of the benign ones (n=11, 68.7%) were hyperintense (P<.001).

Conclusion

DW imaging can be used for reliable discrimination of benign and malignant pediatric abdominal mass lesions based on considerable differences in the ADC values and signal intensity changes.     

Impact of diffusion-weighted MR imaging on the characterization of small hepatocellular carcinoma in the cirrhotic liver

Purpose

The purpose of this study was to determine whether or not adding diffusion-weighted magnetic resonance imaging (DWI) to conventional magnetic resonance (MR) imaging sequences improves the characterization of small hepatocellular carcinoma (HCC) (≤2 cm) in the setting of cirrhotic liver compared to conventional sequences alone.

Materials and Methods

A total of 62 cirrhotic liver patients with 82 nodules smaller than 2 cm in diameter were enrolled, and all lesions were pathologically confirmed. For the first reading session, which included precontrast T1- and T2-weighted images and T1 dynamic contrast-enhanced images, preindicated lesions by a study coordinator were characterized by two radiologists. They determined the confidence levels in consensus for the presence of small HCC into four grades. In another session, respiratory-triggered diffusion-weighted MR images (b factor=50, 400 and 800 s/mm²) were added to the previously reviewed images, and the same two radiologists again determined the confidence levels. The diagnostic performance of the combined DWI–conventional sequences set and the conventional sequences alone set was evaluated using receiver operating characteristic curves. Sensitivity and specificity values for characterizing small HCCs were also calculated.

Results

The area under the receiver operating characteristic curve for the second interpretation session (0.86) was significantly higher (P=.038) than that of the first session (0.76). The sensitivity was significantly increased from 75.7% to 87.8% by adding DWI to the conventional sequences (P=.015). No significant differences were observed for specificity values.

Conclusion

Adding DWI to conventional imaging modalities improves the diagnosis of small HCCs in the cirrhotic liver in terms of diagnostic performance and sensitivity by increasing reader confidence.     

Synovial sarcoma: MR imaging

Ten patients with biopsy-proved synovial sarcoma were evaluated by magnetic resonance (MR) imaging on a 1.5-T unit. The lesions showed intermediate signal intensity on T1-weighted images and heterogeneous high signal intensity on T2-weighted images. Tumors were well-demarcated from normal tissues. Additional information included adjacent bone involvement (one case), femoral vein invasion by tumor (one case), and hemorrhage within the tumors (one case). Four patients underwent a repeat MR examination following chemotherapy. This showed a decrease in size and increase in the signal intensity of three tumors on T2-weighted images, proven to be due to necrosis in one. These changes correlated with clinical regression of disease. While MR in synovial sarcoma does not have any specific signal intensity, it proved to be useful in defining the extent of disease and in determining the response to chemotherapy.     

Handcrafted fuzzy rules for tissue classification

This article proposes a handcrafted fuzzy rule-based system for segmentation and identification of different tissue types in magnetic resonance (MR) brain images. The proposed fuzzy system uses a combination of histogram and spatial neighborhood-based features. The intensity variation from one type of tissue to another is gradual at the boundaries due to the inherent nature of the MR signal (MR physics). A fuzzy rule-based approach is expected to better handle these variations and variability in features corresponding to different types of tissues. The proposed segmentation is tested to classify the pixels of the T2-weighted axial MR images of the brain into three primary tissue types: white matter, gray matter and cerebral-spinal fluid. The results are compared with those from manual segmentation by an expert, demonstrating good agreement between them.     

Quantitative MRI of uterine leiomyomas during triptorelin treatment: Reproducibility of volume assessment and predictability of treatment response

Magnetic resonance (MR) imaging is increasingly applied for the quantitative evaluation of uterine leiomyomas. MR is thought to be more accurate in comparison to ultrasound (US) techniques. MR signal intensity (SI) may prove to be predictive of myoma response to GnRH agonist treatment. This study aimed to evaluate the precision of uterine volume assessment by a parallel planimetric MR method and the accuracy of the ellipsoid formula based calculations from MR and US images. It was also attempted to analyze the precision of MR leiomyoma volume measurements and examine the relation between pretreatment myoma SI patterns and the response to agonist therapy. Twenty-seven women with a myomatous uterus were scanned three times during GnRH agonist treatment for 6 months. T₁- and T₂-weighted, as well as T₁ contrast-enhanced sequences of the uterus were obtained in the transverse and sagittal plane. Abdominal US of the uterus was performed with a conventional sector scanner. By the use of a software system for analysis of three-dimensional images obtained by MR, uterine volume was measured by a parallel planimetric method (MR-ROI) as well as the use of the ellipsoid formula (MR-ELL). Myoma volume was assessed by the MR-ROI method. SI of the myomas was estimated from selected tissue samples as well as from the integral myoma region of interest. By abdominal US, volume was assessed by the ellipsoid equation (US-ELL). Within- and between-observer and method reliability (Rw/Rb) was calculated from mean squares obtained by analysis of variance. For uterine volume assessment, reliability between observers and between methods when the MR-ROI and MR-ELL methods were analyzed was excellent. For the US-ELL measurements, the between-observer reliability was limited. Moreover, the reliability of the US-ELL was low when the MR-ROI method was used as the standard. Myoma volume assessment with the MR-ROI method showed high between-observer and between-method agreement. The myoma/fat SI ratio and the mean SI coefficient of variation failed to show a correlation with the degree of response to triptorelin treatment of individual myomas. In MR uterine volume assessment the MR-ELL method is very accurate compared with the more complicated MR-ROI method. The agreement between MR and US is limited. Therefore, the ellipsoid method on MR images is to be regarded as the method of choice for quantitative assessment of uterine volume response to hormonal treatment. Myoma SI patterns were shown to be of no value in the response prediction of myomas to treatment with GnRH agonists.     

Image enhancement via Median-Mean Based Sub-Image-Clipped Histogram Equalization

This paper presents a robust contrast enhancement algorithm based on histogram equalization methods named Median-Mean Based Sub-Image-Clipped Histogram Equalization (MMSICHE). The proposed algorithm undergoes three steps: (i) The Median and Mean brightness values of the image are calculated. (ii) The histogram is clipped using a plateau limit set as the median of the occupied intensity. (iii) The clipped histogram is first bisected based on median intensity then further divided into four sub images based on individual mean intensity, subsequently performing histogram equalization for each sub image. This method achieves multi objective of preserving brightness as well as image information content (entropy) along with control over enhancement rate, which in turn suits for consumer electronics applications. This method avoids excessive enhancement and produces images with natural enhancement. The simulation results show that MMSICHE method outperforms other HE methods in terms of various image quality measures, i.e. average luminance, average information content (entropy), absolute mean brightness error (AMBE) and background gray level.     

Enhancement of histological volumes through averaging and their use for the analysis of magnetic resonance images

Magnetic resonance imaging (MRI) of small animals is routinely performed in research centers. But despite its many advantages, MR still suffers from limited spatial resolution which makes the interpretation and quantitative analysis of the images difficult, particularly for small structures of interest within areas of significant heterogeneity. One possibility to address this issue is to complement the MR images with histological data, which requires reconstructing 3D volumes from a series of 2D images. A number of methods have been proposed recently in the literature to address this issue, but deformation or tearing during the slicing process often produces reconstructed volumes with visible artifacts and imperfections. In this paper, we show that a possible solution to this problem is to work with several histological volumes, reconstruct each of these separately and then compute an average. The resulting histological atlas shows structures and substructures more clearly than any individual volume. We also propose an original approach to normalize intensity values across slices, a required preprocessing step when reconstructing histological volumes. We show that the histological atlas we have created can be used to localize structures and substructures, which cannot be seen easily in MR images. We also create an MR atlas that is associated with the histological atlas. We show that using the histological volumes to create the MR atlas is better than using the MR volumes only. Finally, we validate our approach quantitatively on MR image volumes by comparing volumetric measurements obtained manually and obtained automatically with our atlases.     

Quantitative mapping of transverse relaxivity (1/T(2)) in hepatic iron overload: a single spin-echo imaging methodology

Recent research into the non-invasive assessment of hepatic iron concentrations using magnetic resonance imaging has shown that the proton transverse relaxivity (1/T(2)) varies linearly with liver iron concentration. However, the development of an image-based system for the assessment of hepatic iron distribution has been confounded by the presence of motion induced artifacts in the T(2)-weighted images. We report on the development of a single spin-echo imaging methodology that enables the generation of transverse relaxivity maps over the liver. A simple smoothing technique is used to accommodate the image intensity perturbations caused by abdominal motion. The relaxivity maps are consistent with the variation of iron concentration throughout the liver. A Parzen density estimate and histogram of the relaxivity distribution are generated to assist in the visual assessment of the degree and variability of T(2) shortening with liver iron loading. It was found that one or two Gaussian functions could be used to characterize the relaxivity distributions with a small number of parameters. We propose that this methodology may be used in the clinical setting to monitor hepatic iron concentrations in the advent of an accurate transverse relaxivity calibration curve.     

结合视觉注意力机制和图像锐度的无参图像质量评价方法

针对模糊图像的质量评价,提出一种新的无参图像质量评价方法,该方法结合了自底向上的视觉注意力机制和自顶向下的图像锐度评价标准。根据人眼视觉注意力机制模型,分别计算颜色、亮度和方向显著度图像,通过竞争机制得到人眼优先关注的区域; 利用无参图像锐度评价方法分别对优先关注的区域及背景区域进行评价,综合2个区域的评价结果得到最终的图像质量评价指标。利用该方法分别对相向运动过程中所产生的模糊图像和图像质量评价Live数据库中的高斯模糊图像进行了评价,结果表明:针对两类图像的评价结果与主观评价结果的相关系数均较高,其中,针对相向运动模糊图像的主客观评价结果的相关系数达到0.98。该方法能够胜任对模糊图像的客观质量评价。     

Analysis of Glaucoma Diagnosis with Automated Classifiers using Stratus Optical Coherence Tomography

The study compared the performances of two classification methods including logistic regression analysis and artificial neural network (ANN) in terms of the area under the receiver operating characteristic curves for differentiating glaucomatous from normal eyes in Taiwan Chinese population based solely on the quantitative assessment of summary data reports from the Stratus optical coherence tomography (OCT). The logistic regression analysis and ANNs showed promise for increasing diagnostic accuracy of glaucoma using summary data from Stratus OCT. The results can be used as the basis for further improving the diagnostic accuracy of glaucoma     

Spatial-Temporal Sensitivity Analysis of Flood Control Capability in China Based on MADM-GIS Model

To facilitate better implementation of flood control and risk mitigation strategies, a model for evaluating the flood defense capability of China is proposed in this study. First, nine indicators such as slope and precipitation intensity are extracted from four aspects: objective inclusiveness, subjective prevention, etc. Secondly, the entropy weight method in the multi-attribute decision making (MADM) model and the improved three-dimensional technique for order preference by similarity to ideal solution (3D-TOPSIS) method were combined to construct a flood defense capacity index evaluation system. Finally, the receiver operating characteristic (ROC) curve and the Taylor plot method were innovatively used to test the model and indicators. The results show that nationwide, there is fine flood defense performance in Shandong, Jiangsu and room for improvement in Guangxi, Chongqing, Tibet and Qinghai. The good representativity of nine indicators selected by the model was verified by the Taylor plot. Simultaneously, the ROC calculated area under the curve (AUC) was 70%, which proved the good problem-solving ability of the MADM-GIS model. An accurate assessment of the sensitivity of flood control capacity in China was achieved, and it is suitable for situations where data is scarce or discontinuous. It provided scientific reference value for the planning and implementation of China’s flood defense and disaster reduction projects and emergency safety strategies.     

Are signal intensity and homogeneity useful parameters for distinguishing between benign and malignant soft tissue masses on MR images?: Objective evaluation by means of texture analysis

Objectives

To objectively identify possible differences in the signal characteristics of benign and malignant soft tissue masses (STM) on magnetic resonance (MR) images by means of texture analysis and to determine the value of these differences for computer-assisted lesion classification.

Method

Fifty-eight patients with histologically proven STM (benign, n=30; malignant, n=28) were included. STM texture was analyzed on routine T1-weighted, T2-weighted and short tau inversion recovery (STIR) images obtained with heterogeneous acquisition protocols. Fisher coefficients (F) and the probability of classification error and average correlation coefficients (POE+ACC) were calculated to identify the most discriminative texture features for separation of benign and malignant STM. F>1 indicated adequate discriminative power of texture features. Based on the texture features, computer-assisted classification of the STM by means of k-nearest-neighbor (k-NN) and artificial neural network (ANN) classification was performed, and accuracy, sensitivity and specificity were calculated.

Results

Discriminative power was only adequate for two texture features, derived from the gray-level histogram of the STIR images (first and 10th gray-level percentiles). Accordingly, the best results of STM classification were achieved using texture information from STIR images, with an accuracy of 75.0% (sensitivity, 71.4%; specificity, 78.3%) for the k-NN classifier, and an accuracy of 90.5% (sensitivity, 91.1%; specificity, 90.0%) for the ANN classifier.

Conclusion

Texture analysis revealed only small differences in the signal characteristics of benign and malignant STM on routine MR images. Computer-assisted pattern recognition algorithms may aid in the characterization of STM, but more data is necessary to confirm their clinical value.     

A novel content-based active contour model for brain tumor segmentation

Brain tumor segmentation is a crucial step in surgical and treatment planning. Intensity-based active contour models such as gradient vector flow (GVF), magneto static active contour (MAC) and fluid vector flow (FVF) have been proposed to segment homogeneous objects/tumors in medical images. In this study, extensive experiments are done to analyze the performance of intensity-based techniques for homogeneous tumors on brain magnetic resonance (MR) images. The analysis shows that the state-of-art methods fail to segment homogeneous tumors against similar background or when these tumors show partial diversity toward the background. They also have preconvergence problem in case of false edges/saddle points. However, the presence of weak edges and diffused edges (due to edema around the tumor) leads to oversegmentation by intensity-based techniques. Therefore, the proposed method content-based active contour (CBAC) uses both intensity and texture information present within the active contour to overcome above-stated problems capturing large range in an image. It also proposes a novel use of Gray-Level Co-occurrence Matrix to define texture space for tumor segmentation. The effectiveness of this method is tested on two different real data sets (55 patients - more than 600 images) containing five different types of homogeneous, heterogeneous, diffused tumors and synthetic images (non-MR benchmark images). Remarkable results are obtained in segmenting homogeneous tumors of uniform intensity, complex content heterogeneous, diffused tumors on MR images (T1-weighted, postcontrast T1-weighted and T2-weighted) and synthetic images (non-MR benchmark images of varying intensity, texture, noise content and false edges). Further, tumor volume is efficiently extracted from 2-dimensional slices and is named as 2.5-dimensional segmentation.     

Synthetic T1-weighted brain image generation with incorporated coil intensity correction using DESPOT1

The increased use of phased-array and surface coils in magnetic resonance imaging, the push toward increased field strength and the need for standardized imaging across multiple sites during clinical trials have resulted in the need for methods that can ensure consistency of intensity both within the image and across multiple subjects/sites. Here, we describe a means of addressing these concerns through an extension of the rapid T(1) mapping technique - driven equilibrium single-pulse observation of T(1). The effectiveness of the proposed approach was evaluated using human brain T(1) maps acquired at 1.5 T with a multichannel phased-array coil. Corrected "synthetic" T(1)-weighted images were reconstructed by substituting the T(1) values back into the governing signal intensity equation while assuming a constant value for the equilibrium magnetization. To demonstrate signal normalization across a longitudinal study, we calculated synthetic T(1)-weighted images from data acquired from the same healthy subject at four different time points. Signal intensity profiles between the acquired and synthetic images were compared to determine the improvements with our proposed approach. Following correction, the images demonstrate obvious qualitative improvement with increased signal uniformity across the image. Near-perfect signal normalization was also observed across the longitudinal study, allowing direct comparison between the images. In addition, we observe an increase in contrast-to-noise ratio (compared with regular T(1)-weighted images) for synthetic images created, assuming uniform proton density throughout the volume. The proposed approach permits rapid correction for signal intensity inhomogeneity without significantly lengthening exam time or reducing image signal-to-noise ratio. This technique also provides a robust method for signal normalization, which is useful in multicenter longitudinal MR studies of disease progression, and allows the user to reconstruct T(1)-weighted images with arbitrary T(1) weighting.     

Magnetic resonance differentiation between T2 and T1 gallbladder carcinoma: significance of subserosal enhancement on the delayed phase dynamic study

Purpose

The aim of this study is to investigate whether subserosal enhancement on the delayed-phase dynamic magnetic resonance (MR) study (SED) can differentiate T2 from T1 gallbladder carcinoma (GBC).

Methods

The institutional research board approved this retrospective study. Between 1997 and 2006, there were surgically proven 11 T1 and 21 T2 GBC in 30 patients, all of whom had undergone preoperative contrast enhanced dynamic MR study, either with a 2D sequence (n=17) or 3D sequences (n=15). All images were reviewed by two radiologists for the presence of SED, and receiver operating characteristic (ROC) curve analysis was performed. Sensitivity, specificity, positive and negative predictive values were calculated by consensus.

Results

The areas under the ROC curves of the two readers were 0.91 and 0.86, and the kappa value was 0.78. Of the 21 T2 GBC, 18 and 3 showed positive and negative SED, respectively. Of the 11 T1 GBC, 1 and 10 showed positive and negative SED, respectively. The sensitivity, specificity, positive and negative predictive values of SED for diagnosing T2 lesions were 86%, 88%, 91% and 77%, respectively.

Conclusions

In conclusion, SED may be a useful sign to differentiate T2 from T1 GBC, which would affect the preoperative surgical planning of the patients.     

基于遗传算法参数优化的PCNN红外图像分割

构造一种基于遗传算法参数优化的脉冲耦合神经网络(PCNN)红外图像分割算法。该算法首先利用PCNN的全局耦合性和脉冲同步性对输入图像进行点火处理,根据PCNN的输出结果计算熵作为遗传算法的适应度函数,并利用熵的变化量作为遗传算法的收敛依据,对PCNN模型中影响图像分割的参数进行组合优化,结合PCNN生物视觉特性和遗传算法解空间随机搜索能力来寻找关键参数的最优值。将遗传算法和PCNN进行结合可充分发挥二者优势,将本文方法与最大类间方差法(OTSU)、最大熵直方图分割算法和PCNN分割方法进行对比,通过交叉熵、区域对比度等客观指标对分割后的图像进行定量分析,结果表明无论从主观视觉还是客观指标,本文方法分割效果优于其他对比方法。