Paulo César Cortez

A New Solution for Automatic Microstructures Analysis from Images Based on a Backpropagation Artificial Neural Network

This article presents a new solution to segment and quantify the microstructures from images of nodular, grey, and malleable cast irons, based on an artificial neural network. The neural network topology used is the multilayer perception, and the algorithm chosen for its training was the backpropagation. This solution was applied to 60 samples of cast iron images and results were very similar to the ones obtained by visual human tests. This was better than the information obtained from a commercial system that is very popular in this area. In fact, this solution segmented the images of microstructures materials more efficiently. Thus, we can conclude that it is a valid and adequate option for researchers, engineers, specialists, and professionals from materials science field to realise a microstructure analysis from images faster and automatically.

Automatic 3D pulmonary nodule detection in CT images

This work presents a systematic review of techniques for the 3D automatic detection of pulmonary nodules in computerized-tomography (CT) images. Its main goals are to analyze the latest technology being used for the development of computational diagnostic tools to assist in the acquisition, storage and, mainly, processing and analysis of the biomedical data. Also, this work identifies the progress made, so far, evaluates the challenges to be overcome and provides an analysis of future prospects. As far as the authors know, this is the first time that a review is devoted exclusively to automated 3D techniques for the detection of pulmonary nodules from lung CT images, which makes this work of noteworthy value. The research covered the published works in the Web of Science, PubMed, Science Direct and IEEEXplore up to December 2014. Each work found that referred to automated 3D segmentation of the lungs was individually analyzed to identify its objective, methodology and results. Based on the analysis of the selected works, several studies were seen to be useful for the construction of medical diagnostic aid tools. However, there are certain aspects that still require attention such as increasing algorithm sensitivity, reducing the number of false positives, improving and optimizing the algorithm detection of different kinds of nodules with different sizes and shapes and, finally, the ability to integrate with the Electronic Medical Record Systems and Picture Archiving and Communication Systems. Based on this analysis, we can say that further research is needed to develop current techniques and that new algorithms are needed to overcome the identified drawbacks.

An innovative approach of QRS segmentation based on first-derivative, Hilbert and Wavelet Transforms

The QRS detection and segmentation processes constitute the first stages of a greater process, e.g., electrocardiogram (ECG) feature extraction. Their accuracy is a prerequisite to a satisfactory performance of the P and T wave segmentation, and also to the reliability of the heart rate variability analysis. This work presents an innovative approach of QRS detection and segmentation and the detailed results of the proposed algorithm based on First-Derivative, Hilbert and Wavelet Transforms, adaptive threshold and an approach of surface indicator. The method combines the adaptive threshold, Hilbert and Wavelet Transforms techniques, avoiding the whole ECG signal preprocessing. After each QRS detection, the computation of an indicator related to the area covered by the QRS complex envelope provides the detection of the QRS onset and offset. The QRS detection proposed technique is evaluated based on the well-known MIT-BIH Arrhythmia and QT databases, obtaining the average sensitivity of 99.15% and the positive predictability of 99.18% for the first database, and 99.75% and 99.65%, respectively, for the second one. The QRS segmentation approach is evaluated on the annotated QT database with the average segmentation errors of 2.85±9.90ms and 2.83±12.26ms for QRS onset and offset, respectively. Those results demonstrate the accuracy of the developed algorithm for a wide variety of QRS morphology and the adaptation of the algorithm parameters to the existing QRS morphological variations within a single record.

Color texture classification using shortest paths in graphs.

Color textures are among the most important visual attributes in image analysis. This paper presents a novel method to analyze color textures by modeling a color image as a graph in two different and complementary manners (each color channel separately and the three color channels altogether) and by obtaining statistical moments from the shortest paths between specific vertices of this graph. Such an approach allows to create a set of feature vectors, which were extracted from VisTex, USPTex, and TC00013 color texture databases. The best classification results were 99.07%, 96.85%, and 91.54% (LDA with leave-one-out), 87.62%, 66.71%, and 88.06% (1NN with holdout), and 98.62%, 96.16%, and 91.34% (LDA with holdout) of success rate (percentage of samples correctly classified) for these three databases, respectively. These results prove that the proposed approach is a powerful tool for color texture analysis to be explored.

Novel and powerful 3D adaptive crisp active contour method applied in the segmentation of CT lung images

&NA; The World Health Organization estimates that 300 million people have asthma, 210 million people have Chronic Obstructive Pulmonary Disease (COPD), and, according to WHO, COPD will become the third major cause of death worldwide in 2030. Computational Vision systems are commonly used in pulmonology to address the task of image segmentation, which is essential for accurate medical diagnoses. Segmentation defines the regions of the lungs in CT images of the thorax that must be further analyzed by the system or by a specialist physician. This work proposes a novel and powerful technique named 3D Adaptive Crisp Active Contour Method (3D ACACM) for the segmentation of CT lung images. The method starts with a sphere within the lung to be segmented that is deformed by forces acting on it towards the lung borders. This process is performed iteratively in order to minimize an energy function associated with the 3D deformable model used. In the experimental assessment, the 3D ACACM is compared against three approaches commonly used in this field: the automatic 3D Region Growing, the level‐set algorithm based on coherent propagation and the semi‐automatic segmentation by an expert using the 3D OsiriX toolbox. When applied to 40 CT scans of the chest the 3D ACACM had an average F‐measure of 99.22%, revealing its superiority and competency to segment lungs in CT images. HighlightsThe 3D Adaptive Crisp Active Contour Method (3D ACACM) is proposed for the segmentation of CT lung images.A new 3D adaptive internal energy is defined to optimize the segmentation.The automatic initialization of the 3D ACACM is described.The 3D ACACM is compared against three methods commonly used in this domain.The results confirm the superior effectiveness of the 3D ACACM. Graphical abstract Figure. No caption available.

Texture analysis and classification using shortest paths in graphs

Texture is a very important attribute in the field of computer vision. This work proposes a novel texture analysis method which is based on graph theory. Basically, we convert the pixels of an image into vertices of an undirected weighted graph and explore the shortest paths between pairs of pixels in different scales and orientations of the image. This procedure is applied to Brodatzs textures and UIUC texture dataset in order to evaluate its capacity of discriminating different kinds of textures. The best classification results using the standard parameters of the method are 98.50%,67.30% and 88.00% of success rate (percentage of samples correctly classified) for Brodatzs textures, UIUC textures (image size of 200x200 pixels), and original UIUC textures (image size of 640x480 pixels), respectively. These results prove that the proposed approach is an efficient tool for texture analysis, once they are superior to the results achieved by traditional and novel texture descriptors presented in literature.

Quantification of the microstructures of hypoeutectic white cast iron using mathematical morphology and an artificial neural network

This paper describes an automatic system for segmentation and quantification of the microstructures of white cast iron. Mathematical morphology algorithms are used to segment the microstructures in the input images, which are later identified and quantified by an artificial neuronal network (ANN). A new computational system was developed because ordinary software could not segment the microstructures of this cast iron correctly, which is composed of cementite, pearlite and ledeburite. For validation purpose, 30 samples were analysed. The microstructures of the material in analysis were adequately segmented and quantified, which did not happen when we used ordinary commercial software. Therefore, the proposed system offers researchers, engineers, specialists and others, a valuable and competent tool for automatic and efficient microstructural analysis from images.

Novel Adaptive Balloon Active Contour Method based on internal force for image segmentation - A systematic evaluation on synthetic and real images

Many studies have been conducted on digital image segmentation, seeking to overcome the limitations of different methods for specific applications. Thus, existing techniques are improved and new methods created. This paper proposes a new Active Contour Method (ACM) applied to the segmentation of objects in digital images. The proposed method is called Adaptive Balloon ACM and its main contribution is the new internal Adaptive Balloon energy that minimizes the energy of each point on the curve using the topology of its neighboring points, and thus moves the curve toward the object of interest. The method can be initialized inside or outside the object of interest, and can even segment objects that have complex shapes. There are no limitations as to its startup location. This work evaluates the proposed method in several applications and compares it with other ACMs in the literature. This new method obtained superior results, especially when the objects to be segmented were tubular and had bifurcations. Thus the proposed method can be considered effective in segmenting complex shapes in digital images and gave promising results in various applications.

New approach for T-wave peak detection and T-wave end location in 12-lead paced ECG signals based on a mathematical model

This paper presents an innovative approach for T-wave peak detection and subsequent T-wave end location in 12-lead paced ECG signals based on a mathematical model of a skewed Gaussian function. Following the stage of QRS segmentation, we establish search windows using a number of the earliest intervals between each QRS offset and subsequent QRS onset. Then, we compute a template based on a Gaussian-function, modified by a mathematical procedure to insert asymmetry, which models the T-wave. Cross-correlation and an approach based on the computation of Trapeziums area are used to locate, respectively, the peak and end point of each T-wave throughout the whole raw ECG signal. For evaluating purposes, we used a database of high resolution 12-lead paced ECG signals, recorded from patients with ischaemic cardiomyopathy (ICM) in the University Hospitals of Leicester NHS Trust, UK, and the well-known QT database. The average T-wave detection rates, sensitivity and positive predictivity, were both equal to 99.12%, for the first database, and, respectively, equal to 99.32% and 99.47%, for QT database. The average time errors computed for T-wave peak and T-wave end locations were, respectively, -0.38±7.12 ms and -3.70±15.46 ms, for the first database, and 1.40±8.99 ms and 2.83±15.27 ms, for QT database. The results demonstrate the accuracy, consistency and robustness of the proposed method for a wide variety of T-wave morphologies studied.

Sistema de segmentação de imagens para quantificação de microestruturas em metais utilizando redes neurais artificiais

Digital Image Processing is an increasing expansion area in some field of application that uses the interpretation of images as tool. Quantitative Metallography area applied to Materials Sciences uses this technique for characterization of phase volumetric fractions, grain size, inclusion distribution determination and other parameters that influence the properties of the materials. The present paper has, as main objective, to present and validate the software Segmentation by Artificial Neural Network (SVRNA), developed by the authors. This software, based on artificial neural network, makes the percentile constituent counting in time reduced in relation to the conventional model. The study is carried out over ABNT 1020 and 1045 steel and nodular cast iron samples. Statistical analysis showed that this software is efficient for admitted degree of significance. It has concluded, therefore, that the program can be used in applications in the field of Material Sciences for determination of microstructures.