Material phase classification by means of Support Vector Machines

Ortegon, Jaime; Ledesma-Alonso, René; Barbosa, Romeli; Castillo, J. Vázquez; Atoche, A. Castillo

doi:10.1016/j.commatsci.2018.02.054

Cited by 23 publications

(7 citation statements)

References 16 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…PCA technique was used for analysis and classification process of bulk metallic glasses materials as a multivariate data analysis method 23 . Support Vector Machine (SVM) method was suggested for material phase classification and has been performed successfully with Otsu-based approach 24 . In a different study, the classification and quantification process of microstructural constituents were investigated using SVM and the parameters affecting to the classification were tested 25 .…”

Section: Introductionmentioning

confidence: 99%

Demonstration of Container Effects on Recognition Process of Liquids Using a Ring-Resonator Measurement Method

Öztürk

2019

Sci Rep

View full text Add to dashboard Cite

Before separating or classifying hazardous/unsafe liquids, a system must be identified which can best measure certain liquids. The preferred measurement system can be useful, quick response, rapid measurement and so on. However, these features are not sufficient to classify the liquids in different containers. Especially at security points, containers of liquids carried by people may be different. Therefore, it has been investigated whether classification can be made by using the response of different containers to electromagnetic radiation. When the ring resonator measurement method is used in the 1–1.4 GHz frequency range, a successful separation process is carried out, even if the containers were made of different materials. Nevertheless, a well-known k-means algorithm has been used to analyze the measurement results of selected liquids.

show abstract

Section: Introductionmentioning

confidence: 99%

Demonstration of Container Effects on Recognition Process of Liquids Using a Ring-Resonator Measurement Method

Öztürk

2019

Sci Rep

View full text Add to dashboard Cite

show abstract

“…Third (SEM1) and fourth (SEM2) images were obtained from SEM. In the case of the SEM image, they are preprocessed to obtain a binary representation using SVM (Support Vector Machine) [4].…”

Section: Resultsmentioning

confidence: 99%

“…Nevertheless, at the microscale, the PEMFC electrode can be modeled by agglomerates in a porous matrix [3], in this scale, a Scanning Electron Microscope (SEM) produces high-resolution images of the microscopic structure. Recently, Ortegon et al presented a classification approach based on Support Vector Machines (SVM) to generate a binarized image from grayscale SEM images of PEMFC electrodes [4]. This heuristic method allows for obtaining a better match with the user’s point of view regarding the solid (agglomerates) and void (porous) phases.…”

Section: Introductionmentioning

confidence: 99%

Effect of An Image Resolution Change on the Effective Transport Coefficient of Heterogeneous Materials

et al. 2019

Self Cite

View full text Add to dashboard Cite

Electrochemical electrodes comprise multiple phenomena at different scales. Several works have tried to model such phenomena using statistical techniques. This paper proposes a novel process to work with reduced size images to reconstruct microstructures with the Simulated Annealing method. Later, using the Finite Volume Method, it is verified the effect of the image resolution on the effective transport coefficient (ETC). The method can be applied to synthetic images or images from the Scanning Electron Microscope. The first stage consists of obtaining the image of minimum size, which contains at least 98% of the statistical information of the original image, allowing an equivalent statistical study. The image size reduction was made by applying an iterative decimation over the image using the normalized coarseness to compare the amount of information contained at each step. Representative improvements, especially in processing time, are achieved by reducing the size of the reconstructed microstructures without affecting their statistical behavior. The process ends computing the conduction efficiency from the microstructures. The simulation results, obtained from two kinds of images from different materials, demonstrate the effectivity of the proposed approach. It is important to remark that the controlled decimation allows a reduction of the processor and memory use during the reconstruction and ETC computation of electrodes.

show abstract

“…These histograms are then used as input features for training the nonlinear SVM. SVM is one of the essential supervised machine learning tools; it has been proposed in many scientific classification fields, such as bioinformatics [38], medical diagnosis [39], environment monitoring [40], and material scientific classification [41]. Designed initially to solve two-class binary classification, SVM has been extended to multiclass classification with two different approaches: One vs. Rest and One vs. One [41,42].…”

Section: The Proposed Methodsmentioning

confidence: 99%

“…SVM is one of the essential supervised machine learning tools; it has been proposed in many scientific classification fields, such as bioinformatics [38], medical diagnosis [39], environment monitoring [40], and material scientific classification [41]. Designed initially to solve two-class binary classification, SVM has been extended to multiclass classification with two different approaches: One vs. Rest and One vs. One [41,42]. SVM uses training data (features) to give the computers acknowledgement without previous programming based on recent advances in statistical learning theory, aiming to maximize the distance between the hyperplane and the support vectors (the samples that effect on the hyperplane) [42].…”

Section: The Proposed Methodsmentioning

confidence: 99%

A Multiscale Topographical Analysis Based on Morphological Information: The HEVC Multiscale Decomposition

et al. 2020

View full text Add to dashboard Cite

In this paper, we evaluate the effect of scale analysis as well as the filtering process on the performances of an original compressed-domain classifier in the field of material surface topographies classification. Each surface profile is multiscale analyzed by using a Gaussian Filter analyzing method to be decomposed into three multiscale filtered image types: Low-pass (LP), Band-pass (BP), and High-pass (HP) filtered versions, respectively. The complete set of filtered image data constitutes the collected database. First, the images are lossless compressed using the state-of-the art High-efficiency video coding (HEVC) video coding standard. Then, the Intra-Prediction Modes Histogram (IPHM) feature descriptor is computed directly in the compressed domain from each HEVC compressed image. Finally, we apply the IPHM feature descriptors as an input of a Support Vector Machine (SVM) classifier. SVM is introduced here to strengthen the performances of the proposed classification system thanks to the powerful properties of machine learning tools. We evaluate the proposed solution we called “HEVC Multiscale Decomposition” (HEVC-MD) on a huge database of nearly 42,000 multiscale topographic images. A simple preliminary version of the algorithm reaches an accuracy of 52%. We increase this accuracy to 70% by using the multiscale analysis of the high-frequency range HP filtered image data sets. Finally, we verify that considering only the highest-scale analysis of low-frequency range LP was more appropriate for classifying our six surface topographies with an accuracy of up to 81%. To compare these new topographical descriptors to those conventionally used, SVM is applied on a set of 34 roughness parameters defined on the International Standard GPS ISO 25178 (Geometrical Product Specification), and one obtains accuracies of 38%, 52%, 65%, and 57% respectively for Sa, multiscale Sa, 34 roughness parameters, and multiscale ones. Compared to conventional roughness descriptors, the HEVC-MD descriptors increase surfaces discrimination from 65% to 81%.

show abstract

Material phase classification by means of Support Vector Machines

Cited by 23 publications

References 16 publications

Demonstration of Container Effects on Recognition Process of Liquids Using a Ring-Resonator Measurement Method

Demonstration of Container Effects on Recognition Process of Liquids Using a Ring-Resonator Measurement Method

Effect of An Image Resolution Change on the Effective Transport Coefficient of Heterogeneous Materials

A Multiscale Topographical Analysis Based on Morphological Information: The HEVC Multiscale Decomposition

Contact Info

Product

Resources

About