Gray-level Histogram Research Articles

Quantitative Radiomics: Impact of Pulse Sequence Parameter Selection on MRI-Based Textural Features of the Brain.

Objectives Radiomic features extracted from diverse MRI modalities have been investigated regarding their predictive and/or prognostic value in a variety of cancers. With the aid of a 3D realistic digital MRI phantom of the brain, the aim of this study was to examine the impact of pulse sequence parameter selection on MRI-based textural parameters of the brain. Methods MR images of the employed digital phantom were realized with SimuBloch, a simulation package made for fast generation of image sequences based on the Bloch equations. Pulse sequences being investigated consisted of spin echo (SE), gradient echo (GRE), spoiled gradient echo (SP-GRE), inversion recovery spin echo (IR-SE), and inversion recovery gradient echo (IR-GRE). Twenty-nine radiomic textural features related, respectively, to gray-level intensity histograms (GLIH), cooccurrence matrices (GLCOM), zone size matrices (GLZSM), and neighborhood difference matrices (GLNDM) were evaluated for the obtained MR realizations, and differences were identified. Results It was found that radiomic features vary considerably among images generated by the five different T1-weighted pulse sequences, and the deviations from those measured on the T1 map vary among features, from a few percent to over 100%. Radiomic features extracted from T1-weighted spin-echo images with TR varying from 360 ms to 620 ms and TE = 3.4 ms showed coefficients of variation (CV) up to 45%, while up to 70%, for T2-weighted spin-echo images with TE varying over the range 60–120 ms and TR = 6400 ms. Conclusion Variability of radiologic textural appearance on MR realizations with respect to the choice of pulse sequence and imaging parameters is feature-dependent and can be substantial. It calls for caution in employing MRI-derived radiomic features especially when pooling imaging data from multiple institutions with intention of correlating with clinical endpoints.

Validation of the Textural Realism of a 3D Anthropomorphic Phantom for Digital Breast Tomosynthesis.

In this paper, texture calculations are used to validate the realism of a physical anthropomorphic phantom for digital breast tomosynthesis. The texture features were compared against clinical mammography data. Three groups of features (grey-level histogram, co-occurrence, and run-length) were considered. The features were analyzed over a broad range of technique settings (kV and mAs). These calculations were done in the central slice of the reconstruction as well as the synthetic 2D mammogram. For each feature, the clinical data were binned into strata based on the compressed breast thickness. It was demonstrated that the clinical features vary by thickness. To evaluate the realism of the phantom, each feature was compared against clinical data in the same thickness stratum. For the purpose of this paper, a feature was considered to be realistic if it was within the middle 95% of the statistical distribution of clinical values. In the reconstruction, most features were found to exhibit realism; specifically, all 12 grey-level histogram features, four out of seven co-occurrence features, and three out of seven run-length features. The realism of most features was robust to changes in the technique settings. However, in the synthetic 2D mammogram, fewer features were found to exhibit realism. In conclusion, this paper provides a validation of the textural realism of the phantom in the reconstruction, and shows that there is less realism in the synthetic 2D mammogram. We identify the features that should be considered to refine the design of the phantom in future work.

Denoising and SAR Image Classification with K-SVD Algorithm

Synthetic Aperture Radar (SAR) is not only having the characteristic of obtaining images during all-day, all-weather, but also provides object information which is distinctive from visible and infrared sensors. but, SAR images have more speckles noise and fewer bands. This paper propose a method for denoising, feature extraction and classification of SAR images. Initially the image was denoised using K-Singular Value Decomposition (K-SVD) algorithm. Then the Gray Level Histogram (GLH) and Gray Level Co-occurrence Matrix (GLCM) are used for extraction of features. Secondly, the extracted feature vectors from the first step were combined using the correlation analysis to decrease the dimensionality of the feature spaces. Thirdly, Classification of SAR images was done in Sparse Representations Classification (SRC) and Support Vector Machines (SVMs). The results indicate that the performance of the introduce SAR classification method is good. The above mentioned classifications techniques are enhanced and performance parameters are computed using MATLAB 2014a software.

A Robust Fuzzy C-Means Algorithm Based on Bayesian Nonlocal Spatial Information for SAR Image Segmentation

The fuzzy c-means (FCM) algorithm and many improved algorithms incorporating spatial information have been proven to be effective in image segmentation. However, these methods are not adaptable to process synthetic aperture radar (SAR) images owing to the intrinsic speckle noise. Our solution, which enables the effective segmentation of SAR images by guaranteeing noise-immunity and edge detail preservation simultaneously, is to propose a robust FCM algorithm based on Bayesian nonlocal spatial information (RFCM $\_$ BNL). The nonlocal idea considers more useful information for generating an auxiliary image. We measure the similarity between patches by utilizing a dedicated noise model for SAR images, and then apply it to the Bayesian formulation. Then we derive a new statistical distance, which is insensitive to speckle noise. Additionally, we ensure that the algorithm is robust to outliers by employing the entropy of the local gray-level histogram to control the extent to which the nonlocal spatial information term is adaptive to pixels. Experiments on simulated and real SAR images show that RFCM $\_$ BNL obtains the best result for SAR image segmentation compared with seven other fuzzy clustering algorithms.

3D bone texture analysis as a potential predictor of radiation-induced insufficiency fractures.

The aim of our work is to assess the potential role of texture analysis (TA), applied to computed tomography (CT) simulation scans, in relation to the development of insufficiency fractures (IFs) in patients undergoing radiation therapy (RT) for pelvic malignancies. We analyzed patients undergoing pelvic RT from Jan-2010 to Dec-2016, 31 of whom had developed IFs of the pelvis. We analyzed CT simulation scans using LifeX Software©, and in particular we selected three regions of interest (ROI): L5 body, the sacrum and both the femoral heads. The ROI were automatically contoured using the treatment planning software Raystation©. TA parameters included parameters from the gray-level histogram, indices from sphericity and from the matrix of GLCM (gray level co-occurrence matrix). The IFs patients were matched (1:1 ratio) with control patients who had not developed IFs, and were matched for age, sex, type of tumor, menopausal status, RT dose and use of chemotherapy. Univariate and multivariate analyses (logistic regression) were used for statistical analysis. Significant TA parameters on univariate analysis included both parameters from the histogram distribution, as well from the matrix of GLCM. On logistic regression analysis the significant parameters were L5-energy [P=0.033, odds ratio (OR): 1.997, 95% CI: 1.059-3.767] and FH-Skewness (P=0.014, OR: 2.338, 95% CI: 1.191-4.591), with a R2: 0.268. A ROC curve was generated from the binary logistic regression, and the AUC was 0.741 (95% CI: 0.627-0.855, P=0.001, S.E.: 0.058). In our experience, 3D-bone CT TA can be used to stratify the risk of the patients to develop radiation-induced IFs. A prospective study will be conducted to validate these findings.

Texture analysis as a predictor of radiation-induced xerostomia in head and neck patients undergoing IMRT.

Image texture analysis (TA) is a heterogeneity quantifying approach that cannot be appreciated by the naked eye, and early evidence suggests that TA has great potential in the field of oncology. The aim of this study is to evaluate parotid gland texture analysis (TA) combined with formal dosimetry as a factor for predicting severe late xerostomia in patients undergoing radiation therapy for head and neck cancers. We performed a retrospective analysis of patients treated at our Radiation Oncology Unit between January 2010 and December 2015, and selected the patients whose normal dose constraints for the parotid gland (mean dose<26Gy for the bilateral gland) could not be satisfied due to the presence of positive nodes close to the parotid glands. The parotid gland that showed the higher V30 was contoured on CT simulation and analysed with LifeX Software©. TA parameters included features of grey-level co-occurrence matrix (GLCM), neighbourhood grey-level dependence matrix (NGLDM), grey-level run length matrix (GLRLM), grey-level zone length matrix (GLZLM), sphericity, and indices from the grey-level histogram. We performed a univariate and multivariate analysis between all the texture parameters, the volume of the gland, the normal dose parameters (V30 and Mean Dose), and the development of severe chronic xerostomia. Seventy-eight patients were included and 25 (31%) developed chronic xerostomia. The TA parameters correlated with severe chronic xerostomia included V30 (OR 5.63), Dmean (OR 5.71), Kurtosis (OR 0.78), GLCM Correlation (OR 1.34), and RLNU (OR 2.12). The multivariate logistic regression showed a significant correlation between V30 (0.001), GLCM correlation (p: 0.026), RLNU (p: 0.011), and chronic xerostomia (p<0.001, R2:0.664). Xerostomia represents an important cause of morbidity for head and neck cancer survivors after radiation therapy, and in certain cases normal dose constraints cannot be satisfied. Our results seem promising as texture analysis could enhance the normal dose constraints for the prediction of xerostomia.

Performance Analysis of Lung Cancer Diagnosis Algorithms on X-Ray Images

Among several types of cancer, lung cancer is regarded as one of the most common and serious. In this respect, early diagnosis is required and beneficial to reduce mortalities caused by this type of cancer. Such diagnosis is typically performed by doctors through manual examinations on X-Ray images. However, manual examinations are labor extensive and time consuming. In this paper, we conduct a study to analyze the performance of some computer-based lung cancer diagnosis algorithms. The algorithms are built using different feature extraction (gray-level co-occurrence matrix, pixel intensity, histogram and combination of the three) and machine learning (Multi-layer Perceptron and K-Nearest Neighbor) techniques and the performance of each algorithm is compared and analyzed. The result of the study shows that the best performance of lung cancer classification is obtained by the computer algorithm that uses the combined features to characterize lung cancer and subsequently classifies the features using Multi-layer Perceptron.

Underwater Image Enhancement by Adaptive Gray World and Differential Gray-Levels Histogram Equalization

Most underwater images tend to be dominated by a single color cast. This paper presents a solution to remove the color cast and improve the contrast in underwater images. However, afte ...

Evaluation of measurement uncertainties of digital volume correlation applied to laminography data

The effect of different data acquisition parameters on the displacement and strain uncertainties of kinematic fields is assessed via digital volume correlation applied to three-dimensional data sets imaged by synchrotron laminography. The measurement uncertainty is estimated for (1) different materials and associated varying contrast; (2) repeated scans, that is, uncertainty due to the imaging system and three-dimensional reconstruction; (3) rigid body motions between two subsequent scans; (4) rescaling of the gray level histogram during 32-bit floating point to 8-bit integer data conversion; (5) changes in the beam properties, that is, use of a monochromator or an undulator; and (6) changes in camera/detector characteristics. It is found that the amount of image contrast is not the only parameter that controls uncertainties for different materials. Furthermore, the rigid body motion procedure should be preferred over repeated scans as it provides more conservative measurement uncertainty values. The applied 32- to 8-bit conversion procedure, beam tuning and detector characteristics hardly affect the measurement uncertainty.

Automatic classification of cast iron grades using support vector machine

In this study, classification of three grades of cast iron viz., gray, malleable and white, based on their texture is attempted, using Haralick features extracted from gray level co-occurrence matrix (GLCM) and histogram features extracted from local binary pattern (LBP). The features were extracted from three hundred images stored in a database and are utilized to train and test the support vector machine (SVM), to classify microstructures. The experimental results show that LBP based feature extraction achieves high accuracy when compared to GLCM based features in classifying cast iron grades.

Material interface detection based on secondary electron images for focused ion beam machining

A method for interface detection is proposed for focused ion beam (FIB) processes of multilayered targets. As multilayers have emerged as promising structures for nanodevices, the FIB machining of multilayers has become a challenging issue. We proposed material interface detection by monitoring secondary electron (SE) images captured during the FIB process. The average of the gray-levels and the skewness coefficient of gray-level histograms of the SE images were evaluated to recognize endpoints for the FIB processes. The FIB process control with the proposed method was demonstrated by fabricating the nanostructures on the multilayered target without thickness information. It was also demonstrated on a curved surface. Grooves with a desired depth into the target and an aperture as an opening window were precisely fabricated by the FIB process control. The proposed strategy of the FIB process can be used for complex substrates such as curved or flexible targets.

A quantitative image analysis for the cellular cytoskeleton during in vitro tumor growth

The cellular cytoskeleton is a dynamic subcellular structure that can be a marker of key biological phenomena including cell division, organelle movement, shape changes and locomotion during the avascular tumor phase. Little attention is paid to quantify changes in the cytoskeleton while nuclei and cytoplasmic both are present in subcellular microscopic images. In this paper, we proposed a quantitative image analysis method to analyze subcellular cytoskeletal changes using a texture analysis method preceded by segmentation of nuclei, cytoplasm and ruffling regions (area except nuclei and cytoplasm). To test and validate this model we hypothesized that Mammary Serine Protease Inhibitor (maspin) acts as cytoskeleton regulator that mediates cell-extracellular matrix (ECM) adhesion in tumor. Maspin-a tumor suppressor gene shows multiple tumor suppressive properties such as increasing tumor cell apoptosis and reducing migration, proliferation, invasion, and overall tumor metastasis. The proposed method obtained separated ruffling regions from segmentation steps and then adopted gray–level histograms (GLH) and grey-level co-occurrence matrix (GLCM) texture analysis techniques. In order to verify the reliability, the proposed texture analysis method was used to compare the control and maspin expressing cells grown on different ECM components: plastic, collagen I, fibronectin and laminin. The results show that the texture parameters extracted reflect the different cytoskeletal changes. These changes indicate that maspin acts as a regulator of the cell-ECM enhancement process, while it reduces the cell migration. Overall, this paper not only presents a quantitative image analysis approach to analyze subcellular cytoskeletal architectures but also provides a comprehensive tool for the biologist, pathologist, cancer specialist, and computer scientist to understand cellular and subcellular organization of cells. In long term, this method can be extended to be used in live cell tracking in vivo, image informatics based point-of-care expert system and quantification of various complex architectures in organisms.

Automatic seed picking for brachytherapy postimplant validation with 3D CT images.

Postimplant validation is an indispensable part in the brachytherapy technique. It provides the necessary feedback to ensure the quality of operation. The ability to pick implanted seed relates directly to the accuracy of validation. To address it, an automatic approach is proposed for picking implanted brachytherapy seeds in 3D CT images. In order to pick seed configuration (location and orientation) efficiently, the approach starts with the segmentation of seed from CT images using a thresholding filter which based on gray-level histogram. Through the process of filtering and denoising, the touching seed and single seed are classified. The true novelty of this approach is found in the application of the canny edge detection and improved concave points matching algorithm to separate touching seeds. Through the computation of image moments, the seed configuration can be determined efficiently. Finally, two different experiments are designed to verify the performance of the proposed approach: (1) physical phantom with 60 model seeds, and (2) patient data with 16 cases. Through assessment of validated results by a medical physicist, the proposed method exhibited promising results. Experiment on phantom demonstrates that the error of seed location and orientation is within ([Formula: see text])mm and ([Formula: see text])[Formula: see text], respectively. In addition, the most seed location and orientation error is controlled within 0.8mm and 3.5[Formula: see text] in all cases, respectively. The average process time of seed picking is 8.7s per 100 seeds. In this paper, an automatic, efficient and robust approach, performed on CT images, is proposed to determine the implanted seed location as well as orientation in a 3D workspace. Through the experiments with phantom and patient data, this approach also successfully exhibits good performance.

Efficacy of Cone-Beam Computed Tomography in Evaluating Bone Quality for Optimum Implant Treatment Planning.

This study examined (1) if cone-beam computed tomography (CBCT) can determine relative differences in bone mineral density distribution using clinical images of patients' mandibular bone and (2) if the relative differences can be used to detect the effects of sex and age on bone mineral density distribution. Sixty-six clinical CBCT images from patients (36 females and 30 males) of 3 age groups (40, 50, and 60 years) were identified. Alveolar (AB) and basal cortical bone (CB) regions were digitally isolated. A histogram of gray levels, which are proportional to degrees of bone mineralization, was obtained from each region. Mean, variability (SD and coefficient of variation), and percentage differences of gray level parameters between AB and basal CBs were computed. Significant sex differences in gray level variability were observed within the postmenopausal age group (P < 0.042). These findings suggest that clinical CBCT images can be a valuable tool in providing information on bone quality, which is an important criterion for optimum planning for dental implant placement.

An intelligent system for quality measurement of Golden Bleached raisins using two comparative machine learning algorithms

In this research, an expert system is provided for measuring and recognizing the quality and purity of mixed (pure-impure) raisins using bulk raisins’ images. For this purpose, by utilizing a machine vision setup 1400 images of the raisins were captured in the several ranges of mixture (from 5 to 50%). Then, totally 146 textural features were obtained using four methods of gray-level histograms, gray level co-occurrence matrix (GLCM), gray level run-length (GLRM) matrix, and local binary pattern (LBP). Principal Components Analysis (PCA) was used in order to find the optimum features from the extracted features. Accordingly, Artificial Neural Network (ANN) and Support Vector Machine (SVM) were used for classifying the mixtures. In comparison to ANN, using top 50 features, SVM classifier had more efficient and accurate classification results (averagely 92.71%). The results of the proposed approach can be used in designing a system for purity and quality measuring of raisins.

Focus ultrasound for augmenting convection-enhanced delivery of nanoparticles in the brain

We previously demonstrated how ultrasound can enhance the dispersion of locally administrated nanoparticles within the extracellular/perivascular spaces in the ex vivo brain by non-destructively enlarging these regions. The current study aimed to translate these results in vivo, where custom, non-adhering brain-penetrating nanoparticles (BPN: 60, 200, and 500 nm), were administered directly into the brains of Sprague Dawley rats by convection-enhanced delivery. Non-invasive, transcranial focused ultrasound (TCFUS) was carried out using an MRI-guided system (1.5 MHz, 10 ms pulses, 10% duty cycle, and 2.3 MPa). 15 individual exposures in a 3 × 5 matrix (spacing: 1.5 mm) in one hemisphere were given, where the size of the focal zone (-6 dB) was 1 × 1 X 8 mm. At 2hrs post-treatment brains were harvested and sectioned, with digital images captured and processed using a custom MATLAB script. This involved the “Otsu” thresholding method, based on gray level histograms and threshold determinations for maximizing the interclass variance. As expected, BPN distributions in the non-treated brains decreased with an increase in diameter. Pretreating with TCFUS was found to significantly increase the distribution of the 200 nm BPNs. These results have broad implications for therapeutic delivery for a variety of brain diseases and disorders.

A Method for Extracting and Selecting Features of Thermal Infrared Remote Sensing Images of Harbor Targets

To improve the inadequate effectiveness of thermal infrared (TIR) image recognition due to low contrast, striping noise and low spatial resolution, this paper conducted research with respect to harbor targets. It adopted sample features like texture, geometry, shape and statistics, selected the best combination of image features with an attribute selection evaluator, and chose the best classifier and the best classifier parameters according to the accuracy of recognition. This process can generate the best features for the classification of TIR images with a certain degree of robustness. The results of the experiment indicate that 1) the libSVM (a support vector machine), after parameter optimization, is the most accurate classifier and the best result can reach 82.4137%; 2) the one-dimensional statistical features of pixel value, mean value, variance, mid-value, and gray level histogram are more meaningful for the recognition; and 3) daytime images can be more accurately classified than nighttime ones.

Gray Level Histogram Width Tissue Characterization

Purpose Proposal to study liver with gray level histogram width GLHW clinical ultrasound tissue characterization using usual B mode devices Materials Placental intervillous space fibrin deposit in fetal growth restriction FGR malignant neoplasia fetal brain immature fetal lung meconium stained amniotic fluid and healthy adult liver Methods Commercial ultrasound B mode histogram base width divided by full gray scale length was gray level histogram width GLHW and it was used to tissue characterization Results GLHW of ovarian cervical and endometrial malignancies were or higher Fibrin deposit in placental intervillous space was diagnosed with GLHW and treated with heparin administration to obtain normal neonate Fetal brain echo density immature fetal lung meconium stained amniotic fluid and healthy adult liver GLHW were studied with GLHW Helsinki declaration was followed Conclusion The GLHW objectively diagnosed ultrasound B mode image thus it would be applied even in adult human cases

Use of High-Resolution X-Ray Micro-Computed Tomography for the Analysis of Internal Structural Changes in Maize Infected with Fusarium verticillioides

X-ray micro-computed tomography (X-ray micro-CT) is a non-destructive, three-dimensional (3D) imaging and analysis technique for the investigation of internal structure of a large variety of materials, including agricultural produce. As a relatively new method in the field of food science, X-ray micro-CT has been applied successfully to obtain micro-structural information of foods undergoing different physical and chemical changes. In this study, high-resolution X-ray micro-CT was used for non-destructive analysis of the internal structure of maize kernels infected with Fusarium verticillioides. The major anatomical features of the maize kernel were identified based on their differences in X-ray attenuation, i.e. the germ, scutellum, vitreous and floury endosperm. Fungal infection caused changes in the internal structure of the kernels over time, which included a decrease in total kernel volume and an increase in total volume of void space, with more voids observed in the germ and floury endosperm regions. No significant (P > 0.05) difference was observed between the control and the infected kernels; it was apparent that the changes observed in the infected kernels were not solely as a result of fungal growth. The grey level histograms of the control and infected kernels shifted to the lower grey value intensity range over time indicating an increase in void space within the kernels. In the 3D images, the increase in total volume of void space with fungal progression was clearer and the effect of fungal damage on the internal structure was evident.

A fuzzy fusion approach for modified contrast enhancement based image forensics against attacks

In today’s digital age the trustworthy towards image is distorting because of malicious forgery images. The issues related to the multimedia security have led to the research focus towards tampering detection. The main objective of the work is to develop robust and forensic detection framework against post processing. It is also essential to enhance the security against attacks. In this paper, a Modified Contrast Enhancement based Forensics (MCEF) method based on Fuzzy Fusion is proposed against post-processing activity. First, we check for the histogram peaks and gaps as a result of contrast enhancement which is used in the latest technique. From the standpoint of attackers, we use two types of attacks, CE trace hiding attack and CE trace forging attack, which could invalidate the forensic detector and fabricate two types of forensic errors, consequently. The CE trace hiding attack is implemented by integrating local random dithering into the form of pixel value mapping. The CE trace forging attack is proposed by modifying the grey level histogram of a target pixel region to fraudulent peak/gap artifacts. Then both attacks are added to enhanced images as a post processing activity. As a result the gaps get disappeared, but introduced sudden peaks. Then, feature selection methods in conjunction with fuzzy fusion approach is suggested to enhance the robustness of tamper detection methods. The threshold value for contrast detection is increased, so we can identify the contrast enhancement. The Artificial Neural Network (ANN) is used instead of SVM, it increases the robustness and accuracy of the digital images. The proposed methodology will be implemented using MATLAB and validated by comparing with the conventional techniques.