Automatic microstructural characterization and classification using probabilistic neural network on ultrasound signals

Vejdannik, Masoud; Sadr, Ali

doi:10.1007/s10845-016-1225-y

Cited by 11 publications

(4 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…SVM is also far slower than this, and demands more computational requirement than OPF. On the other hand, the proposed neural networks in (Vejdannik and Sadr 2016b) and (Vejdannik and Sadr 2016c) provide much higher accuracy and speed. In these neural networks, four different approaches are presented for estimating the probability density function.…”

Section: 5classification Methodsmentioning

confidence: 99%

“…This implies the high correlation between the adjacent samples, because each sample is followed by another sample with an almost identical amplitude. In contrast, samples that are spaced at 20-point spacing are less correlated, because data points are randomly distributed in the scattering plot (Vejdannik and Sadr 2016a, Vejdannik and Sadr 2016b, Vejdannik and Sadr 2016c.…”

Section: Applications Of Signal Processing To Ndtmentioning

confidence: 99%

“…While the proper combination of methods, such as hybrid neuro-fuzzy networks, can provide an appropriate and efficient solution. For example, Vejdannik and Sadr in (Vejdannik and Sadr 2016b) and (Vejdannik and Sadr 2016c) optimized the PNN (Probabilistic Neural Network) and RBF (Radial Basis Function) neural networks using the Bees Algorithm and so they were able to improve the accuracy of classifying ultrasound signals for characterizing the microstructures within the materials.…”

Section: Soft Computing Hard Computingmentioning

confidence: 99%

“…In addition, in some applications of the NDT, we seek to characterize the physical and mechanical properties of materials. For example, Vejdannik and Sadr in (Vejdannik and Sadr 2016a, Vejdannik and Sadr 2016b, Vejdannik and Sadr 2016c processed and classified the backscattered ultrasound signals which have been acquired from the specimens of nickel alloy, in order to investigate the microstructural properties of these materials. Certainly, this application is not feasible in the traditional way, or at least scanning electron microscopes should be used, which is time-consuming and costly.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Signal Processing for NDE

Vejdannik¹,

Sadr²,

Albuquerque³

et al. 2018

Handbook of Advanced Non-Destructive Evaluation

Self Cite

View full text Add to dashboard Cite

Nowadays, testing and evaluating of industrial equipment using nondestructive tests, is a fundamental step in the manufacturing process. The complexity and high costs of manufacturing industrial components, require examinations in some way about the quality and reliability of the specimens. However, it should be noted, that in order to accurately perform the nondestructive test, in addition to theoretical knowledge, it is also essential to have the experience and carefulness, which requires special courses and experience with theoretical education. Therefore, in the traditional methods, which are based on manual testing techniques and the test results depend on the operator, there is the possibility of an invalid inference from the test data. In other words, the accuracy of conclusion from the obtained data is dependent on the skill and experience of the operator. Thus, using the signal processing techniques for nondestructive evaluation (NDE), it is possible to optimize the methods of nondestructive inspection, and in other words, to improve the overall system performance, in terms of reliability and system implementation costs. In recent years, intelligent signal processing techniques have had a significant impact on the progress of nondestructive assessment. In other words, by automating the processing of nondestructive data and signals, and using the artificial intelligence methods, it is possible to optimize nondestructive inspection methods. Hence, improve overall system performance in terms of reliability and Implementation costs of the system. This chapter reviews the issues of intelligent processing of nondestructive testing (NDT) signals.

show abstract

Section: 5classification Methodsmentioning

confidence: 99%

Section: Applications Of Signal Processing To Ndtmentioning

confidence: 99%