Tool For Texture Analysis Research Articles

Weak to no correlation between quantitative high-resolution computed tomography metrics and lung function change in fibrotic diseases.

Identifying systemic sclerosis (SSc) and idiopathic pulmonary fibrosis (IPF) patients at risk of more rapid forced vital capacity (FVC) decline could improve trial design. The purpose of the present study was to explore the prognostic value of quantitative high-resolution computed tomography (HRCT) metrics derived by Imbio lung texture analysis (LTA) tool in predicting FVC slope. This retrospective study used data from patients who were not treated with investigational drugs with and without background antifibrotic therapies in tocilizumab phase 3 SSc, lebrikizumab phase 2 IPF, and zinpentraxin alfa phase 2 IPF studies conducted from 2015 to 2021. Controlled HRCT axial volumetric multidetector computed tomography scans were evaluated using the Imbio LTA tool. Associations between HRCT metrics and FVC slope were assessed through the Spearman correlation coefficient and adjusted R2 in a linear regression model adjusted by demographics and baseline clinical characteristics. A total of 271 SSc and IPF patients were analysed. Correlation coefficients of highest magnitude were observed in the SSc study between the extent of ground glass, normal volume, quantification of interstitial lung disease, reticular pattern, and FVC slope (-0.25, 0.28, -0.28, and -0.33, respectively), while the correlation coefficients observed in IPF studies were in general <0.2. The incremental prognostic value of the baseline HRCT metrics was marginal after adjusting baseline characteristics and was inconsistent across study arms. Data from the SSc and IPF studies suggested weak to no and inconsistent correlation between quantitative HRCT metrics derived by the Imbio LTA tool and FVC slope in the studied SSc and IPF population.

Can CT texture analysis parameters be used as imaging biomarkers for prediction of malignancy in canine splenic tumors?

Splenic hemangiosarcoma has morphological similarities to benign nodular hyperplasia. Computed tomography (CT) texture analysis can analyze the texture of images that the naive human eye cannot detect. Recently, there have been attempts to incorporate CT texture analysis with artificial intelligence in human medicine. This retrospective, analytical design study aimed to assess the feasibility of CT texture analysis in splenic masses and investigate predictive biomarkers of splenic hemangiosarcoma in dogs. Parameters for dogs with hemangiosarcoma and nodular hyperplasia were compared, and an independent parameter that could differentiate between them was selected. Discriminant analysis was performed to assess the ability to discriminate the two splenic masses and compare the relative importance of the parameters. A total of 23 dogs were sampled, including 16 splenic nodular hyperplasia and seven hemangiosarcoma. In each dog, total 38 radiomic parameters were extracted from first-, second-, and higher-order matrices. Thirteen parameters had significant differences between hemangiosarcoma and nodular hyperplasia. Skewness in the first-order matrix and GLRLM_LGRE and GLZLM_ZLNU in the second, higher-order matrix were determined as independent parameters. A discriminant equation consisting of skewness, GLZLM_LGZE, and GLZLM_ZLNU was derived, and the cross-validation verification result showed an accuracy of 95.7%. Skewness was the most influential parameter for the discrimination of the two masses. The study results supported using CT texture analysis to help differentiate hemangiosarcoma from nodular hyperplasia in dogs. This new diagnostic approach can be used for developing future machine learning-based texture analysis tools.

Evaluation of Erosion Potential and Determination of The Causes of Erosion Using A Geographic Information System

The purpose of this study was to determine the magnitude of the erosion class and too determine the cause distribution area in the Upper Serang Sub-watershed. In this study using combinatian of quantitative and qualitative descriptive methods. While the tools used are GPS, soil sampling tools, laboratory test equipment such as permeability analysis tools, organic materials tools, texture analysis tools, data analysis tools using Arc Gis 10.4.1 and SPSS. Erosion potential analysis using USLE (Universal Soil Loss Equation) method. Based on the research results, the Upper Serang Sub-watershed has 5 erosion class categories, i,e very light class (0.5-10.11 t·ha–1·y–1). Light grade (15.26-48.52 t·ha–1·y–1). Medium class (70.55-164.51 t·ha–1·y–1). Heavy class (208.95–388.41 t·ha–1·y–1). Very heavy class (697.47–2241.14 t·ha–1·y–1). While the Erosion Hazard Index is divided into 4 classes, i,e low class with a value of &lt;1. Medium class with a score of 1.001-4.00. High class with a value of 4.01–10.0. Very high grade with a score of &gt; 10. The dominant factor influencing the high and low erosion value is the slope factor and the length of the slope at the location of the Upper Serang Sub-watershed based on the closeness value with a value of 0.935 with a very strong closeness class. Keywords: Erosion, USLE, Serang Upper, Sub-watershed

Green fluorescent protein and phase contrast image fusion via Spectral TV filter-based decomposition

The spectral filter is one of the crucial tools for signal scale and texture analysis. The main function is to decompose a raw image into a series of multi-scale signatures in the spectral domain. For cross-modal fusion, most of the existing approaches naively employ the decomposition of spectral features to achieve image reconstruction. Unfortunately, it does not sufficiently explore the potential cues of different level components in the spatial domain. In this work, we describe an effective decomposition approach using the spectral filter. Specifically, we aim to build a multi-scale image decomposition framework from the spectral domain to the spatial domain, which is called the multi-scale spatial decomposition approach (MSD). Moreover, building upon spectral total variation (TV), we characterize the spectral signatures into various spatial levels with different filters. A fusion scheme of green fluorescent protein and phase contrast image is introduced, leveraging the multi-scale spatial decomposition approach. The source image is first separated by applying MSD into three spatial bands, namely high-frequency bands, low-frequency bands and structure bands. Our method is simple and effective. The divided bands are obtained by a set of filters and could be processed with the specific fusion strategies to extract the features of the source image. In addition, experiments show that our method produces promising results than other current popular methods.

Automated surface texture analysis via Discrete Cosine Transform and Discrete Wavelet Transform

Surface roughness and texture are critical to the functional performance of engineering components. The ability to analyze roughness and texture effectively and efficiently is much needed to ensure surface quality in many surface generation processes, such as machining, surface mechanical treatment, etc. Discrete Wavelet Transform (DWT) and Discrete Cosine Transform (DCT) are two commonly used signal decomposition tools for surface roughness and texture analysis. Both methods require selecting a threshold to decompose a given surface into its three main components: form, waviness, and roughness. However, although DWT and DCT are part of the ISO surface finish standards, there exists no systematic guidance on how to compute these thresholds, and they are often manually selected on case by case basis. This makes utilizing these methods for studying surfaces dependent on the user's judgment and limits their automation potential. Therefore, we present two automatic threshold selection algorithms based on information theory and signal energy. We use machine learning to validate the success of our algorithms both using simulated surfaces as well as digital microscopy images of machined surfaces. Specifically, we generate feature vectors for each surface area or profile and apply supervised classification. Comparing our results with the heuristic threshold selection approach shows good agreement with mean accuracies as high as 95%. We also compare our results with Gaussian filtering (GF), and show that while GF results for areas can yield slightly higher accuracies, our results outperform GF for surface profiles. We further show that our automatic threshold selection has significant advantages in terms of computational time as evidenced by decreasing the number of mode computations by an order of magnitude compared to the heuristic thresholding for DCT.

Chronic lung allograft dysfunction phenotype and prognosis by machine learning CT analysis.

Chronic lung allograft dysfunction (CLAD) is the principal cause of graft failure in lung transplant recipients and prognosis depends on CLAD phenotype. We used a machine learning computed tomography (CT) lung texture analysis tool at CLAD diagnosis for phenotyping and prognostication compared with radiologist scoring. This retrospective study included all adult first double lung transplant patients (January 2010-December 2015) with CLAD (censored December 2019) and inspiratory CT near CLAD diagnosis. The machine learning tool quantified ground-glass opacity, reticulation, hyperlucent lung and pulmonary vessel volume (PVV). Two radiologists scored for ground-glass opacity, reticulation, consolidation, pleural effusion, air trapping and bronchiectasis. Receiver operating characteristic curve analysis was used to evaluate the diagnostic performance of machine learning and radiologist for CLAD phenotype. Multivariable Cox proportional hazards regression analysis for allograft survival controlled for age, sex, native lung disease, cytomegalovirus serostatus and CLAD phenotype. 88 patients were included (57 bronchiolitis obliterans syndrome (BOS), 20 restrictive allograft syndrome (RAS)/mixed and 11 unclassified/undefined) with CT a median 9.5 days from CLAD onset. Radiologist and machine learning parameters phenotyped RAS/mixed with PVV as the strongest indicator (area under the curve (AUC) 0.85). Machine learning hyperlucent lung phenotyped BOS using only inspiratory CT (AUC 0.76). Radiologist and machine learning parameters predicted graft failure in the multivariable analysis, best with PVV (hazard ratio 1.23, 95% CI 1.05-1.44; p=0.01). Machine learning discriminated between CLAD phenotypes on CT. Both radiologist and machine learning scoring were associated with graft failure, independent of CLAD phenotype. PVV, unique to machine learning, was the strongest in phenotyping and prognostication.

Multiscale permutation entropy for two-dimensional patterns

Complexity measures are important to understand and analyze systems with one dimensional data. However, extension of these methods to images (two dimensional data) are much less usual. Bidimensional multiscale sample entropy (MSE2D) has recently been proposed as a new complexity measure for texture evaluation. However, MSE2D leads to undefined or unreliable values for small-sized textures and requires a long computation time. This is why we herein propose the bidimensional multiscale permutation entropy (MPE2D) to evaluate the complexity of 2D patterns. MPE2D is applied to different synthesized textures, to softwood samples, and to study the texture of breast histopathology images. The results show that MPE2D is a valuable tool for texture analysis and that it is computationally noticeably faster than MSE2D.

Open Source Filament Diameter Sensor for Recycling, Winding, and Additive Manufacturing Machines

Abstract To overcome the challenge of upcycling plastic waste into three-dimensional (3D) printing filament in the distributed recycling and additive manufacturing systems, this study designs, builds, tests, and validates an open-source filament diameter sensor for recycling and winding machines. The modular system for multi-axis optical control of the diameter of the recycled 3D-printer filament makes it possible to scan part of the surface of the processed filament, save the history of measurements along the entire length of the spool, as well as mark defective areas. The sensor is developed as an independent module and integrated into a recyclebot. It was tested on different kinds of polymers (acrylonitrile butadiene styrene (ABS), polylactide (PLA)), different sources of plastic, and different colors including clear plastic. The results were compared with the manual measurements, and the measurements obtained with a one-dimensional digital light caliper. The results found that the developed open-source filament sensing method allows users to obtain significantly more information in comparison with basic one-dimensional light sensors and using the received data not only for more accurate diameter measurements but also for a detailed analysis of the recycled filament surface. This could help to expand the use of plastic recycling technologies in the manufacturing community. The availability of tools for possible texture analysis could also stimulate the growth of composite materials creation. The presented system can greatly enhance the user possibilities and serve as a starting point for a complete recycling control system that will regulate motor parameters to achieve the desired filament diameter with acceptable deviations and even control the extrusion rate on a printer to recover from filament irregularities.

Kinematic Fields Measurement during Orthogonal Cutting Using Digital Images Correlation: A Review

Understanding the mechanism of chip formation during orthogonal cutting requires a local measurement of the displacement and strain fields in the cutting zone. These measurements can then be used in order to enhance/validate numerical simulation of metal cutting or calibrate material behavior laws for a better prediction of the thermomechanical loads inside the cutting zone. Particle tracking to identify the strain localization that is exhibited in the Adiabatic Shear Band (ASB) is a challenging task. These local measurements can be determined by images post-processing while using the Digital Image Correlation (DIC) technique or analytical models using streamline models or by micro grid analysis. Recently, the use of the DIC technique is widely increased. Texture quality has been shown to be an important factor. Various techniques of surface preparation are then discussed and classified in terms of the created pattern size. Tools for texture analysis are presented. The technique suitability for the kinematic fields measurement while using the DIC technique during machining is discussed. Various optical systems of the literature employed in the context of kinematic fields measurement during machining are discussed in this paper. The recent advances on the design of optical systems are given. Finally, the results of kinematic fields measurement during machining metallic alloys are analyzed.

Combined texture analysis and machine learning in suspicious calcifications detected by mammography: Potential to avoid unnecessary stereotactical biopsies

ObjectivesTo investigate whether combined texture analysis and machine learning can distinguish malignant from benign suspicious mammographic calcifications, to find an exploratory rule-out criterion to potentially avoid unnecessary benign biopsies. MethodsMagnification views of 235 patients which underwent vacuum-assisted biopsy of suspicious calcifications (BI-RADS 4) during a two-year period were retrospectively analyzed using the texture analysis tool MaZda (Version 4.6). Microcalcifications were manually segmented and analyzed by two readers, resulting in 249 image features from gray-value histogram, gray-level co-occurrence and run-length matrices. After feature reduction with principal component analysis (PCA), a multilayer perceptron (MLP) artificial neural network was trained using histological results as the reference standard. For training and testing of this model, the dataset was split into 70 % and 30 %. ROC analysis was used to calculate diagnostic performance indices. Results226 patients (150 benign, 76 malignant) were included in the final analysis due to missing data in 9 cases. Feature selection yielded nine image features for MLP training. Area under the ROC-curve in the testing dataset (n = 54) was 0.82 (95 %-CI: 0.70−0.94) and 0.832 (95 %-CI 0.72−0.94) for both readers, respectively. A high sensitivity threshold criterion was identified in the training dataset and successfully applied to the testing dataset, demonstrating the potential to avoid 37.1–45.7 % of unnecessary biopsies at the cost of one false-negative for each reader. ConclusionCombined texture analysis and machine learning could be used for risk stratification in suspicious mammographic calcifications. At low costs in terms of false-negatives, unnecessary biopsies could be avoided.

TAU AND AMYLOID PATHOLOGY IN ASSOCIATION WITH SUBJECTIVE COGNITIVE PERFORMANCE IN NORMAL ELDERLY AND EARLY MILD COGNITIVE IMPAIRMENT

Introduction Subjective cognitive decline (SCD) is increasingly recognized as the initial outward sign of preclinical Alzheimer's disease. The ability to explore associations between individual domains (memory, language, executive function, and visuospatial processing) within SCD and pathological biomarkers of early Alzheimer's disease (AD) would provide an important tool to understand the pathological basis of SCD and the risk of developing AD. This study utilizes multiple data analysis strategies to characterize the severity of tau and amyloid burden in clinically normal elderly adults (CN) and early mild cognitive impairment (EMCI) subjects in association with their subjective cognitive decline. Methods All subject data were downloaded from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database. These included [18F]flortaucipir tau and [18F]florbetapir amyloid positron emission tomography (PET) images as well as the subjects’ SCD scores. For SCD, we used the self- and informant-reported individual items of Everyday Cognition questionnaire, including Everyday Memory, Everyday Language, Everyday Visuospatial abilities, Everyday Planning, Everyday Organization, and Everyday Divided Attention. The severity of the amyloid burden was calculated as the global PET standardized uptake value ratio (SUVR). The severity of the tau burden was assessed with both standard uptake-based approaches (tau SUVR in entorhinal, limbic, and isocortical Braak regions) and with a novel texture analysis tool, the weighted two-point correlation (wS2) analysis, which characterizes the spatial clustering of the tau-PET image as a sensitive early biomarker of abnormal tau accumulation and spread. Using a series of backward-elimination regression models, we identified the combination of predictors (from tau PET, global amyloid-PET, sex, APOE, and age) that constructed the best model to predict each Everyday Cognition item. Results Across the different cognitive tests, the wS2-based tau measures explained more variability, as indicated by p-values and elevated R-squared values, than standardized uptake value ratios from different anatomical regions, including the entorhinal cortex which is believed to be among the first regions to show signs of tau pathology. Our backward-elimination regression analyses (figure 1) suggested that the strongest models for predicting some of the subjective cognitive items (Everyday Memory, Everyday Language, Everyday Organization) were more likely amyloid-related whereas other items (Everyday Visuospatial, Everyday Planning) were more tau-related (together with age and sex). Conclusions Different pathological pathways may influence the manifestation of different subjective cognitive items. In vivo texture analysis techniques may be used as sensitive tools to explore these pathological pathways in association with the earliest signs of cognitive changes.

Radiomics in RayPlus: a Web-Based Tool for Texture Analysis in Medical Images.

Radiomics has been shown to have considerable potential and value in quantifying the tumor phenotype and predicting the treatment response. In most scenarios, the commercial and open-source software programs are available for quantitative analysis in medical images to streamline radiomics research. However, at this stage, most of these programs are local applications and require users to have experience in programming and software engineering, which clinicians usually do not have. Therefore, in this article, a web-based tool was proposed to flexibly support radiomics research workflow tasks. Radiomics in RayPlus requires zero installation, is easy to maintain, and accessible anywhere via any PC or MAC with an Internet connection. The system provides functions including multimodality image import and viewing, ROI definition, feature extraction, and data sharing. As a web application, it appears an effective way to multi-institution and multi-department collaborative radiomics research and moreover, its transparency, flexibility, and portability can greatly accelerate the pace of clinical data analysis.

Comparing the Reliability of Biomedical Texture Analysis Tools on Different Image Types

Comparing the Reliability of Biomedical Texture Analysis Tools on Different Image Types

Variable angle gray level co-occurrence matrix analysis of T2 relaxation time maps reveals degenerative changes of cartilage in knee osteoarthritis: Oulu knee osteoarthritis study.

Texture analysis methods based on gray level co-occurrence matrices (GLCM) can be optimized to probe the spatial correspondence information from knee MRI T2 maps and the changes caused by osteoarthritis, and thus possibly leading to a more sensitive characterization of osteoarthritic cartilage. Curvature of the cartilage surfaces combined with the low image resolution in relation to cartilage thickness set special requirements for an effective texture analysis tool. To introduce a novel implementation of GLCM algorithm optimized for cartilage texture analysis; to evaluate the performance of the designed algorithm against mean T2 relaxation time analysis; and to identify the most suitable texture features for discerning osteoarthritic subjects and asymptomatic controls. Case control. Eighty symptomatic osteoarthritis patients and 64 asymptomatic controls. Multislice multiecho spin echo sequence on a 3T MRI system. The T2 relaxation time maps were manually segmented by an operator trained for the task. Texture analysis was performed using an in-house algorithm developed in MATLAB. Symptomatic and asymptomatic subjects were compared using Mann-Whitney U-test. Repeatability of different features was addressed using the concordance correlation coefficient (CCC). Spearman's correlations between the features were determined. The algorithm displayed excellent performance in discerning symptomatic and asymptomatic subjects. Fifteen features provided a significant difference between the groups (P ≤ 0.05) and 12 of those had P values smaller than the mean T2 differences. Most of the studied texture features were highly repeatable with CCC over 90%. For the features with significant differences, correlation with mean T2 was low or moderate (|r| ≤ 0.5). With careful parameter and feature selection and algorithm optimization, texture analysis provides a powerful tool for assessing knee osteoarthritis with more sensitive detection of cartilage degeneration compared to the mean value of the T2 relaxation times in an identical region of interest. 2 Technical Efficacy Stage 2 J. Magn. Reson. Imaging 2018;47:1316-1327.

Texture analysis of the developing human brain using customization of a knowledge-based system

Background:Pattern recognition software originally designed for geospatial and other technical applications could be trained by physicians and used as texture analysis tools for evidence-based practice, in order to improve diagnostic imaging examination during pregnancy.Methods:Various machine-learning techniques and customized datasets were assessed for training of an integrable knowledge-based system (KBS) to determine a hypothetical methodology for texture classification of closely related anatomical structures in fetal brain magnetic resonance (MR) images. Samples were manually categorized according to the magnetic field of the MRI scanner (i.e., 1.5-tesla [1.5T], 3-tesla [3T]), rotational planes (i.e., coronal, sagittal, and axial), and signal weighting (i.e., spin-lattice, spin-spin, relaxation, and proton density). In the machine-learning sessions, the operator manually selected relevant regions of interest (ROI) in 1.5/3T MR images. Semi-automatic procedures in MaZda/B11 were performed to determine optimal parameter sets for ROI classification. Four classes were defined: ventricles, thalamus, gray matter, and white matter. Various texture analysis methods were tested. The KBS performed automatic data preprocessing and semi-automatic classification of ROI.Results:After testing 3456 ROI, statistical binary classification revealed that the combination of reduction techniques with linear discriminant algorithms (LDA) or nonlinear discriminant algorithms (NDA) yielded the best scoring in terms of sensitivity (both 100%, 95% CI: 99.79–100), specificity (both 100%, 95% CI: 99.79–100), and Fisher coefficient (≈E+4 and ≈E+5, respectively).Conclusions:LDA and NDA in MaZda can be useful data mining tools for screening a population of interest subjected to a clinical test.

Texture Characterization Using Shape Co-Occurrence Patterns.

Texture characterization is a key problem in image understanding and pattern recognition. In this paper, we present a flexible shape-based texture representation using shape co-occurrence patterns. More precisely, texture images are first represented by a tree of shapes, each of which is associated with several geometrical and radiometric attributes. Then, four typical kinds of shape co-occurrence patterns based on the hierarchical relationships among the shapes in the tree are learned as codewords. Three different coding methods are investigated for learning the codewords, which can be used to encode any given texture image into a descriptive vector. In contrast with existing works, the proposed approach not only inherits the shape-based method's strong ability to capture geometrical aspects of textures and high robustness to variations in imaging conditions but also provides a flexible way to consider shape relationships and to compute high-order statistics on the tree. To the best of our knowledge, this is the first time that co-occurrence patterns of explicit shapes have been used as a tool for texture analysis. Experiments on various texture and scene data sets demonstrate the efficiency of the proposed approach.

CT Texture Analysis of Ex Vivo Renal Stones Predicts Ease of Fragmentation with Shockwave Lithotripsy.

Understanding the factors affecting success of extracorporeal shockwave lithotripsy (SWL) would improve informed decision-making on the most appropriate treatment modality for an individual patient. Although stone size and skin-to-stone distance do correlate with fragmentation efficacy, it has been shown that stone composition and architecture, as reflected by structural heterogeneity on CT, are also important factors. This study aims to determine if CT texture analysis (CTTA), a novel, nondestructive, and objective tool that generates statistical metrics reflecting stone heterogeneity, could have utility in predicting likelihood of SWL success. Seven spontaneously passed, intact renal tract stones, were scanned ex vivo using standard CT KUB and micro-CT. The stones were then fragmented in vitro using a clinical lithotripter, after which, chemical composition analysis was performed. CTTA was used to generate a number of metrics that were correlated to the number of shocks needed to fragment the stone. CTTA metrics reflected stone characteristics and composition, and predicted ease of SWL fragmentation. The strongest correlation with number of shocks required to fragment the stone was mean Hounsfield unit (HU) density (r = 0.806, p = 0.028) and a CTTA metric measuring the entropy of the pixel distribution of the stone image (r = 0.804, p = 0.039). Using multiple linear regression analysis, the best model showed that CTTA metrics of entropy and kurtosis could predict 92% of the outcome of number of shocks needed to fragment the stone. This was superior to using stone volume or density. CTTA metrics entropy and kurtosis have been shown in this experimental ex vivo setting to strongly predict fragmentation by SWL. This warrants further investigation in a larger clinical study for the contribution of CT textural metrics as a measure of stone heterogeneity, along with other known clinical factors, to predict likelihood of SWL success.

LBP maps for improving fractal based texture classification

This paper presents an innovative manner of obtaining discriminative texture signatures by using the LBP approach to extract additional sources of information from an input image and by using fractal dimension to calculate features from these sources. Four strategies, called Min, Max, Diff Min and Diff Max, were tested, and the best success rates were obtained when all of them were employed together, resulting in an accuracy of 99.25%, 72.50% and 86.52% for the Brodatz, UIUC and USPTex databases, respectively, using Linear Discriminant Analysis. These results surpassed all the compared methods in almost all the tests and, therefore, confirm that the proposed approach is an effective tool for texture analysis.

Upper and lower volumetric fractal descriptors for texture classification

In this work, we propose a novel set of fractal descriptors extracted from gray-level texture images using Bouligand–Minkowski method. To accomplish that, we first convert the texture image into a surface, where the z-axis is the intensity value associated to a pixel in the image. Then, we use a sphere of radius r to dilate this surface. However, instead of using only the influence volume information to extract the descriptors, we considered that the arrangement of the pixels in the texture interferes in the dilation process differently in regions of concavity and convexity of the texture. Thus, we separate these descriptors in two sets: upper and lower volumetric fractal descriptors. Classification results of the descriptors show that they presents a great discrimination power as they overcome traditional methods found in literature in three different benchmark databases, proving to be an efficient tool for texture analysis.

Alloying propagation in nanometric Ni/Al multilayers: A molecular dynamics study

In nanometric metallic multilayers such as Ni/Al, the alloying reaction proceeds in the form of a propagating wave. We studied the different phase transformations involved in the reactive wave propagation by means of molecular dynamics. The focus was on a specific regime that involves melting of reactants, intermixing of reactants, and formation of an intermetallic compound. We found that the wave consists of two stages. The first front is associated with a dissolution process and propagates at several meters per second, while the second front is due to the crystallization of the final product and is slower, leading to a specific microstructure with alternated large grains of NiAl and liquid regions in the front propagation direction. Three main exothermic processes were identified, including grain coarsening. Their respective contributions were evaluated. We developed a new texture analysis tool that allowed us to follow the evolution of the microstructure and the dynamics of the grain orientation.