Gray-level Images Research Articles

Iris-based Image Processing for Cholesterol Level Detection using Gray Level Co-Occurrence Matrix and Support Vector Machine

Serious illnesses such as strokes and heart attacks can be triggered by high levels of cholesterol in human blood that exceeds ideal conditions, where the ideal cholesterol level is below 200 mg/dL. To find out cholesterol levels need a long process because the patient must go through a blood sugar test that requires the patient to undergo fasting for 10–12 hours first before the test. Iridology is a branch of science that studies human iris and its relation to the wellness of human internal organs. The method can be used as an alternative for medical analysis. Iridology thus can be used to assess the conditions of organs, body construction, and other psychological conditions. This paper proposes a cholesterol detection system based on the iris image processing using Gray Level Co-Occurrence Matrix (GLCM) and Support Vector Machine (SVM). GLCM is used as the feature extraction method of the image, while SVM acts as the classifier of the features. In addition to GLCM and SVM, this paper also construct a preprocessing method which consist of image resizing, segmentation, and color image to gray level conversion of the iris image. These steps are necessary before the GLCM feature extraction step can be applied. In principle, the GLCM method is a construction of a matrix containing the information about the proximity position of gray level images pixels. The output of GLCM is fed to the SVM that relies on the best hyperplane. Thus, SVM performs as a separator of two data classes of the input space. From the simulation results, the system built was able to detect excess cholesterol levels through iris image and classify into three classes, namely: non–cholesterol (< 200 ), risk of cholesterol (200–239 ) and high cholesterol (> 240 ). The accuracy rate obtained was 94.67% with an average computation time of 0.0696 . It was using each of the 75 training and test data, with the second-order parameters used are contrast–correlation–energy–homogeneity, pixel distance = 1, quantization level = 8, Polynomial kernel types and One Against One Multiclass. Iris has specific advantages which can record all organ conditions, body construction and psychological conditions. Therefore, Iridology as a science based on the arrangement of iris fibers can be an alternative for medical analysis. In this paper proposes a cholesterol detection system through the iris using Gray Level Co-occurence Matrix and Support Vector Machine. Input system is an iris image that will be processed by pre-processing and then extracted features with the Gray Level Co-Occurrence Matrix method which is a matrix containing information about position the proximity of pixels that have a certain gray level. And then classified with the Support Vector Machine method that relies on the best hyper lane which functions as a separator of two data classes in the input space. From the simulation results, the system built was able to detect excess cholesterol levels through iris image and classify into three classes are: risk of cholesterol, high cholesterol and non–cholesterol with an accuracy rate of 96.47% and average computation time was 0.0696 using each of the 75 training and test data, with the second-order parameters used are contrast–correlation–energy–homogeneity, pixel distance = 1, quantization level = 8, Polynomial kernel types and One Against One Multiclass.

Equivalence between Digital Well-Composedness and Well-Composedness in the Sense of Alexandrov on n-D Cubical Grids

Among the different flavors of well-composednesses on cubical grids, two of them, called, respectively, digital well-composedness (DWCness) and well-composedness in the sense of Alexandrov (AWCness), are known to be equivalent in 2D and in 3D. The former means that a cubical set does not contain critical configurations, while the latter means that the boundary of a cubical set is made of a disjoint union of discrete surfaces. In this paper, we prove that this equivalence holds in n-D, which is of interest because today images are not only 2D or 3D but also 4D and beyond. The main benefit of this proof is that the topological properties available for AWC sets, mainly their separation properties, are also true for DWC sets, and the properties of DWC sets are also true for AWC sets: an Euler number locally computable, equivalent connectivities from a local or global point of view. This result is also true for gray-level images thanks to cross section topology, which means that the sets of shapes of DWC gray-level images make a tree like the ones of AWC gray-level images.

Gray-level image denoising with an improved weighted sparse coding

The nonlocal self-similarity of images means that groups of similar patches have low-dimensional property. The property has been previously used for image denoising, with particularly notable success via sparse coding. However, only a few studies have focused on the varying statistics of noise in different similar patches during the iterative denoising process. This has motivated us to introduce an improved weighted sparse coding for gray-level image denoising in this paper. On the basis of traditional sparse coding, we introduce a weight matrix to account for the noise variation characteristics of different similar patches, while introduce another weight matrix to make full use of the sparsity priors of natural images. The Maximum A-Posterior estimation (MAP) is used to obtain the closed-form solution of the proposed method. Experimental results demonstrate the competitiveness of the proposed method compared with that of state-of-the-art methods in both the objective and perceptual quality.

Phase contrast tomography to study near-field effects of polypropylene fibres on hardened cement paste

In this study the boundary zone around polypropylene fibres in hardened cement paste was defined in terms of porosity and unhydrated cement particles using the phase contrast tomography with a spatial resolution of 0.637 μm within a volume of 1.305×1.305×1.305 mm3. Pores and unhydrated grains were identified from the cement paste made of blended cement by separating them from computed tomography images using thresholding method. In the identification of the gray-level ranges of pores and unhydrated particles several approaches, such as tangent-slope and overflow criterion including the cumulative curve of the gray-level image, were utilised. The changes in relative amounts of pores and unhydrated particles indicated that the boundary zone extended to a distance of 70 μm from the fibre surface. Within the boundary zone, the porosity was about twice or even four times higher than in the bulk cement paste. The volume fraction of pores was observed to vary from 7% to 35% close to the fibre, where the pores were also typically interconnected with the resolution used. Compared to bulk cement paste, the amount of pores exceeding a volume of 105 μm3 increased in the boundary zone. Close to the fibre surface, the volume fraction of unhydrated particles was about 2% being 33% of the value outside the boundary zone.

A Control Circuit Design for Automatic Aperture

In view of the fact that automatic aperture technique is applied as a method of regulating light in the current industrial monitoring and control systems more and more, an automatic aperture control circuit is designed according to automatic aperture control principle. The circuit has the characteristics of the simple structure, low cost and high reliability and etc. It has been proved by experiments that the circuit can adjust the size of aperture automatically according to the change of the external illumination, which ensures proper luminous flux entering the lens, and avoids the loss of image gray levels in bright space, and solves the image flicker caused by the change of light intensity, ensures the CCD get used to brightness changes of scenery outside and obtain the best image.

Remote sensing image colorization using symmetrical multi-scale DCGAN in YUV color space

Image colorization technique is used to colorize the gray-level image or single-channel image, which is a very significant and challenging task in image processing, especially the colorization of remote sensing images. This paper proposes a new method for coloring remote sensing images based on deep convolution generation adversarial network. The adopted generator model is a symmetrical structure using the principle of auto-encoder, and a multi-scale convolutional module is specially designed to introduce into the generator model. Thus, the proposed generator can enable the whole model to retain more image features in the process of up-sampling and down-sampling. Meanwhile, the discriminator uses residual neural network 18 that can compete with the generator, so that the generator and discriminator can effectively optimize each other. In the proposed method, the color space transformation technique is first utilized to convert remote sensing images from RGB to YUV. Then, the Y channel (a gray-level image) is used as the input of the neural network model to predict UV channels. Finally, the predicted UV channels are concatenated with the original Y channel as a whole YUV that is then transformed into RGB space to get the final color image. Experiments are conducted to test the performance of different image colorization methods, and the results show that the proposed method has good performance in both visual quality and objective indexes on the colorization of remote sensing image.

On the development of a coupled nonlinear telegraph-diffusion model for image restoration

In this work, we propose a telegraph coupled partial differential equation (TCPDE) based model for image restoration. The new framework interpolates between a couple of nonlinear telegraph equation and a parabolic equation. This proposed strategy can be applied to significantly preserve the oscillatory and texture pattern in an image, even in low signal-to-noise ratio (SNR). First, we prove that the present model has a unique global weak solution using Banach’s fixed point theorem. Then we apply our model over a set of gray level images to illustrate the image denoising ability of the proposed model. The experimental results of the proposed model are reported, which found better in terms of noise elimination and edge preservation, with respect to the recently developed hyperbolic–parabolic PDE based models as well as coupled diffusion-based model.

On the use of autocorrelation functions, permeability tensors, and computed tomography to characterise the anisotropy of diesel particulate filter materials

ABSTRACT We show how the combination of the spatial autocorrelation function and permeability calculations, applied to 3D X-ray computed tomography data, can yield quantitative information on the anisotropy of both meso-structure and fluid flow in Diesel Particulate Filter (DPF) materials, such as Cordierite and SiC. It was found that both the degree of anisotropy, and the orientation of the permeability and meso-structure are similar, but not identical. We confirm that the morphological anisotropy of cordierite materials is weak, and clearly influenced by the extrusion process that determines the main direction of anisotropy. Properties of the autocorrelation function are discussed and it is shown why estimating the characteristic length of real meso-structures (grain or ‘pore’ size) is not possible. Finally, we show that the autocorrelation function applied on grey-level images can give a good estimate of the degree of anisotropy even with limited resolution.

Cropping Method on Grayscale Images for Periapical Radiographs of Human Teeth

This study explains the comparison of techniques for cropping existing in Matlab for graylevel images on periapical radiographs of human teeth. Some cropping techniques in Matlab are rectangular crop, square crop, circular crop, ellipse crop, polygon crop. This research will explain how each existing cropping technique will give different results based on their respective shapes. So that eventually will produce the most appropriate cropping techniques among several existing cropping techniques in mathlab. The five cropping techniques that have been mentioned here where the four existing cropping techniques namely rectangular crop, square crop, circular crop, ellipse crop still carry other objects that are not needed in research, so it is not appropriate to be used for research objects that have irregular shapes. Appropriate cropping technique is polygon cropping because it is able to identify research objects that have irregular shapes.

A study of dispersion in chromatic confocal microscopy using digital image processing

A research about digital image processing in chromatic confocal microscopy for the study of dispersion of a homogeneous semitransparent material is presented. The deduction of a mathematical model to determine the refractive index with a chromatic confocal system is made first. Then, a methodology using digital image processing in conjunction with that mathematical model in order to determine the dispersion is proposed. The main parts of this methodology consisted in making an axial scanning of the sample, filling a matrix with the spectral data acquired, obtaining a 256 gray level image (that we called optical thickness matrix or OTM) from this matrix, applying binarization and skeletonization to this image and carrying out a fitting process to obtain a curve. It was found that by means of either horizontal or vertical cuttings in the OTM it is possible to obtain the dispersion of a material. The vertical cuttings are special because they have an interesting analogy with a hypothetical multiple confocal laser system. The OTM can be a very useful tool because it could show a lot of visual information about the sample. Thus, it could give a fast visual estimate of values and characteristics associated with the sample.

A Machine Learning Model to Detect Flow Disturbances during Manufacturing of Composites by Liquid Moulding

In this work, a supervised machine learning (ML) model was developed to detect flow disturbances caused by the presence of a dissimilar material region in liquid moulding manufacturing of composites. The machine learning model was designed to predict the position, size and relative permeability of an embedded rectangular dissimilar material region through use of only the signals corresponding to an array of pressure sensors evenly distributed on the mould surface. The burden of experimental tests required to train in an efficient manner such predictive models is so high that favours its substitution with synthetically-generated simulation datasets. A regression model based on the use of convolutional neural networks (CNN) was developed and trained with data generated from mould-filling simulations carried out through use of OpenFoam as numerical solver. The evolution of the pressure sensors through the filling time was stored and used as grey-level images containing information regarding the pressure, the sensor location within the mould and filling time. The trained CNN model was able to recognise the presence of a dissimilar material region from the data used as inputs, meeting accuracy expectation in terms of detection. The purpose of this work was to establish a general framework for fully-synthetic-trained machine learning models to address the occurrence of manufacturing disturbances without placing emphasis on its performance, robustness and optimization. Accuracy and model robustness were also addressed in the paper. The effect of noise signals, pressure sensor network size, presence of different shape dissimilar regions, among others, were analysed in detail. The ability of ML models to examine and overcome complex physical and engineering problems such as defects produced during manufacturing of materials and parts is particularly innovative and highly aligned with Industry 4.0 concepts.

A Fuzzy Segmentation Method to Learn Classification of Mitosis

Mitotic counts are widely used as a metric for cellular proliferation for prognosis and to determine the aggressiveness of individual cancers. This study presents a less labor-intensive method to count mitotic cells in breast cell sections. The proposed algorithm involves two phases: candidate segmentation and detection. During candidate segmentation, images are filtered through a blue ratio threshold to remove unnecessary background information and to increase the color difference between targets and non-targets for an entire digitized image. A fuzzy candidate segmentation method is used to adaptively determine threshold values in order to dichotomize gray-level images and distinguish the images of mitotic candidates from the background. The thresholding scheme integrates the spatial characteristics’ distribution in a histogram to determine an intensity threshold for the processed image, in order to filter insignificant information. During the detection phase, a two-class classification uses an attention mechanism that is realized by a set of fully connected neural networks, instead of convolutional layers, which decreases the computational cost. The validation test using ICPR2012 competition datasets shows that the proposed model outperforms current state-of-art techniques, in terms of the metrics, Accuracy, F1-score, and Precision and Recall.

Crop Analysis Using Image Processing

To detect the uniqueness and quantities of agriculture product images a new method is proposed using MATLAB software .In this paper we propose a method to increase the contrast level of a image with exponential low pass filter and histogram equalization technique. Next by using region props function we extract the binary features of the image, and then we calculated the number of targets in gray level image. This method can be easily applied in modern agriculture.

Double stages of feature extarction-based GFPMI for colored finger vein identification

Today, finger vein recognition has a lot of attention as a promising approach of biometric identification framework and still does not meet the challenges of the researchers on this filed. To solve this problem, we propose s double stage of feature extraction schemes based localized finger fine image detection. We propose Globalized Features Pattern Map Indication (GFPMI) to extract the globalized finger vein line features basede on using two generated vein image datasets: original gray level color, globalized finger vein line feature, original localized gray level image, and the colored localized finger vein images. Then, two kinds of features (gray scale and texture features) are extracted, which tell the structure information of the whole finger vein pattern in the whole dataset. The recurrent based residual neural network (RNN) is used to identify the finger vein images. The experimental show that the localized colored finger vein images based globalized feature extraction has achieved the higher accuracy (93.49%) while the original image dataset achieved less accuracy by (69.86%).

Enhancing the Classification of Pap Smear Images using ENN – TLBO classification Method

The most common cancer among the women younger than 35 in developing countries is cervical cancer. It is a human papilloma virus disease. It should be identified earlier by Pap smear test and treated earlier to avoid the consequences. Pap test a colonoscopy is widely used to check the vagina and the cervix. The Pap smear test is the most effective medical test, but it causes difficulty under the microscope at the point of analysis. Automatic cancer detection is designed to unravel the downside. This identification process involves some image processing methods, such as segmentation, and an improved SVM classification method. In this paper, an efficient Elman Neural Network (ENN) collaborating with Teaching Learning Based Optimization (TLBO) algorithm is proposed to classify cancer using Pap Smear Test images. At first, an input image of Pap smear is converted into grey level from the RGB. The grey level image is preprocessed to eliminate unwanted noise produced and smoothened with Kuan Filter (KF). Active Contour Method (ACM) has been used to segment the identified cells from the Pap smear image. Features such as GLCM, haralick, solidity, shape, and other mathematical features are extracted for improving the accuracy. Classification has done using ENN-TLBO. TLBO is utilized for getting optimal weights during the training phase. Performance evaluation has done through the experimental outcomes where, ENN-TLBO yields good accuracy of 86.6%, than the prevailing algorithms such as Support Vector Machine, Radial Basis Function Neural Network classifiers.

Gray Level Image Encoding in Plasmonic Metasurfaces

Plasmonic metasurfaces have been widely used for image and color representation using nanoscale structures. By designing the size or shape of nanostructures, the phase or amplitude of transmitted or reflected electromagnetic spectrum can be manipulated via control of resonant wavelength, which results in multiple colors and can be used for color and/or image generation. Instead of color printing, we introduce a new approach from which images with multiple gray scales can be generated and hidden under light with specific wavelengths. In this work, plasmonic structures of nanorods with identical geometries are fabricated in arrays on a glass substrate, but the nanorods comprise binary alloys of aluminum and titanium; therefore, grayscale image information is encoded with the material composition. Under white light illumination with polarization along the long axis of a nanorod, the response of the nanorods has essentially the same resonant frequency, but with varying magnitudes of transmission. Based on this, using nanorods with four different compositions, four gray level imaging has been achieved. Since nanorods are sensitive to the polarization and spectral composition of the incident light, under unpolarized white light illumination, the image disappears. Therefore, this technique is promising in image encryption. We also show that the manipulation of light at the resonant wavelength by this method occurs in amplitude, not phase.

A NEW APPROACH OF FRACTAL COMPRESSION USING COLOR IMAGE

Fractal image compression depends on self-similarity, where one segment of a image is like the other one segment of a similar picture. Fractal coding is constantly connected to grey level images. The simplest technique to encode a color image by gray- scale fractal image coding algorithm is to part the RGB color image into three Channels - red, green and blue, and compress them independently by regarding each color segment as a specific gray-scale image. The colorimetric association of RGB color pictures is examined through the calculation of the relationship essential of their three-dimensional histogram. For normal color images, as a typical conduct, the connection necessary is found to pursue a power law, with a non- integer exponent type of a given image. This conduct recognizes a fractal or multiscale self-comparable sharing of the colors contained, in average characteristic pictures. This finding of a conceivable fractal structure in the colorimetric association of regular images complement other fractal properties recently saw in their spatial association. Such fractal colorimetric properties might be useful to the characterization and demonstrating of natural images, and may add to advance in vision. The outcomes got demonstrate that the fractal-based compression for the color image fills in similarly with respect to the color image.

Computer-aided detection and morphological characterization of nanotube layers using scanning electron microscopy images

We propose an approach for computer-aided morphological analysis of titania nanotube arrays by using high-resolution scanning electron microscopy images. Based on top-view images and several preset parameters, our experimental MATLAB® routine finds possible candidates for nanotube openings and validates them by size, solidity, and eccentricity. Centers and inner diameters are computed followed by a local validation of pixels belonging to nanotube openings and walls. To achieve this goal, input gray level images are repeatedly converted to black and white using the entire range of thresholds from 254 to 1 to reveal black patches as nanotube openings. Nanotube locations are identified level after level in a systematic manner, taking into account the local gray level distribution. This approach allowed to identify most nanotubes in a reasonable computation time. When the nanotube detection is completed, the routine computes the nanotube surface density, the average nanotube diameter, the percentage of area occupied in the input image by nanotubes and unreacted titania, and the distribution of nanotubes based on their diameter, solidity and eccentricity. Our method provides a reliable and automated quality assessment of titania nanotube array layers used in sensor-based and photocatalytic applications. The total number of the detected nanotubes in the SEM images is dramatically increased by 30% to 40% when compared to previous methods. The proposed procedure can be easily adjusted for morphological analysis of any other materials structured as nanotube layers.

Classification of Brodatz Dataset Samples (36-64) Through Second-Order Statistical Parameters (B)

In this research paper, we have extended the approach of evaluating the texture features of the Brodatz dataset for the samples varying from 36-64. Here we have mathematically analyzed the texture visual features for the Brodatz dataset since we are aware of the fact that contrast, correlation, energy, and homogeneity are considered as the texture features which can affect human vision, so we have calculated the values of the features of texture feature for visual parameters only. As a result, we have obtained information about the samples which has reached the maximum and minimum feature value. Here we have also shown the grey level images of the dataset.

Quantitative imaging of 226Ra ultratrace distribution using digital autoradiography: Case of doped celestines

The ability of the autoradiographic device BeaQuant™ is evaluated herein to quantitatively map the ultratrace element 226Ra distributed spatially in celestine (SrSO4) grains/crystals. 226Ra doped celestines have been obtained from coprecipitation and recrystallization experiments, and have been characterized with high purity germanium gamma detector (HPGe), giving specific activities ranging from 3251 to 32523 Bq.g−1. Alpha autoradiographs of polished sections from doped celestines have been obtained using BeaQuant™. These alpha maps have been compared to the celestine grains/crystals arrangement observed with a scanning electron microscope (SEM). At the sample scale, celestine grains are responsible of an alpha signal, indicating that 226Ra is detectable in celestine from its alpha emissions. 226Ra distribution has also been investigated at the celestine grains/crystals scale: the crystal/grain properties do not allow to decide if the distribution process is homogeneous or not, i.e. if there is a chemical zoning into the crystal/grain. The counting of alpha particles by autoradiography has been compared with the total activity of the 226Ra doped celestines by gamma counting (HPGe technique). This comparison was performed by standardizing the measured activities to the same celestine volume, which has been determined by performing a threshold on SEM grey level images to assess to the celestine surface and using Geant4 Monte Carlo simulation toolkit to assess to the emission depth of the particles in celestine. A very good linear correlation between gamma activity and alpha counting from autoradiographs is obtained for all the samples, demonstrating the ability of BeaQuant™ to quantify 226Ra in any points of the millimetric section samples, at a resolution of 20 μm.