Fault Diagnosis Based on Tree Heuristic Feature Selection and FS-DFV for Rolling Element Bearings DOI Open Access

Xiaoyue chen,

Xiong Liu, Ge Dang

и другие.

IOP Conference Series Materials Science and Engineering, Год журнала: 2019, Номер 630(1), С. 012024 - 012024

Опубликована: Окт. 1, 2019

Abstract In order to make up for the deficiency of traditional single diagnosis in rolling element bearing fault application, eliminate a large amount redundant information and improve classification effect aliasing mode, based on comprehensive analysis respective advantages fuzzy set tree search, this paper presents joint method tree-inspired feature selection FS-DFV (Fuzzy Set Dependent Feature Vector). The dependent vectors (DFV) can dig deeper essential differences faults accuracy. By establishing heuristic model, type search strategy is designed, excellent criteria density clustering with noise are proposed, conventional model improved. addition, used process problem extracting patterns DFV, membership guide subsequent extraction alias modes. proposed compared other four methods. experimental results show that effectively diagnostic efficiency bearing.

Язык: Английский

Research and application of a novel hybrid air quality early-warning system: A case study in China DOI
Chen Li,

Zhijie Zhu

The Science of The Total Environment, Год журнала: 2018, Номер 626, С. 1421 - 1438

Опубликована: Фев. 19, 2018

Язык: Английский

Процитировано

72
Validity and reliability of forensic firearm examiners DOI
Erwin J.A.T. Mattijssen,

Cilia Witteman,

Charles E.H. Berger

и другие.

Forensic Science International, Год журнала: 2019, Номер 307, С. 110112 - 110112

Опубликована: Дек. 18, 2019

Язык: Английский

Процитировано

49
Automated Firearm Classification From Bullet Markings Using Deep Learning DOI Creative Commons

Pattranit Pisantanaroj,

Pimlapus Tanpisuth,

Piyawut Sinchavanwat

и другие.

IEEE Access, Год журнала: 2020, Номер 8, С. 78236 - 78251

Опубликована: Янв. 1, 2020

Firearm violence is one of the leading causes death in many countries around world, including Thailand. This work proposes a fast and accurate automated method to classify firearm brands from bullet markings. Specifically, panoramic image collected crime scene was captured using developed mobile phone application custom-built portable hardware. The top three state-of-the-art CNNs pretrained on ImageNet-DenseNet121, ResNet50, Xception-were further trained same training set, which composed 718 bullets eight different brands-Beretta, Browning, CZ, Glock, Norinco, Ruger, Sig Sauer, Smith & Wesson-using five-fold cross validation technique. DenseNet121 provided highest AUC 0.99 for CZ classification (the most common registered brand Thailand) average (0.9780 ± 0.0130 SD), significantly higher than those ResNet50 Xception. In addition, there were no interaction effects between CNN model AUC. DenseNet121, had AUC, evaluated test set (72 bullets), results showed that Beretta classifications lowest accuracy (91.18%), followed by Browning Norinco (96.88%), whereas Wesson (98.41%). These suggest based deep learning algorithm hardware have promising potential use at scenes firearms By narrowing down list suspects, this convenient approach can potentially accelerate identification processes forensic science examiners.

Язык: Английский

Процитировано

15
Interpol review of forensic firearm examination 2019–2022 DOI Creative Commons
Erwin J.A.T. Mattijssen, Wim Kerkhoff,

Rob Hermsen

и другие.

Forensic Science International Synergy, Год журнала: 2022, Номер 6, С. 100305 - 100305

Опубликована: Дек. 14, 2022

Язык: Английский

Процитировано

6
Objective Identification of Bullets Based on 3D Pattern Matching and Line Counting Scores DOI

Danny Roberge,

A.L. Beauchamp,

Serge Lévesque

и другие.

International Journal of Pattern Recognition and Artificial Intelligence, Год журнала: 2019, Номер 33(11), С. 1940021 - 1940021

Опубликована: Фев. 19, 2019

In firearm identification, a examiner looks at pair of fired bullets or cartridge cases using comparison microscope and determines from this visual analysis if they were both the same firearm. particular case bullets, individual signature takes form striated pattern. Over time, examiner’s community developed two distinct approaches for bullet identification: pattern matching line counting. More recently, emergence technology enabling capture surface topographies down to submicron depth resolution has been catalyst field computerized objective ballistic identification. Objectiveness is achieved through statistical various scores known matches nonmatches exhibit comparison, which in turn implies large quantities topographies. The main goal study was develop an identification method conventionally rifled barrels, test on public proprietary 3D image datasets captured different lateral resolutions. Two newly scores, Line Counting Score (LCS) Pattern Matching Score, computed yielded perfect match versus nonmatch separation three sets used standard Hamby–Brundage Test. A similar performed larger, more-realistic set, enabled us define discriminative false rate 1/10[Formula: see text]000 2D plot that shows nonmatches. LCS shown produce better sensitivity than consecutive striae criteria dataset. likelihood function also linear combination conservative approach based extreme value theory proposed extrapolate score domain where data are not available. This provides understanding limitations studies involve very few firearms.

Язык: Английский

Процитировано

9
Interpol review of forensic firearm examination 2016-2019 DOI Creative Commons
Erwin J.A.T. Mattijssen

Forensic Science International Synergy, Год журнала: 2020, Номер 2, С. 389 - 403

Опубликована: Янв. 1, 2020

This review paper covers the relevant literature on forensic firearm examination from 2016 to 2019 as a part of 19th Interpol International Forensic Science Managers Symposium. The papers are also available at website at: https://www.interpol.int/content/download/14458/file/Interpol%20Review%20Papers%202019.pdf.

Язык: Английский

Процитировано

6
Identification of bullets fired from air guns using machine and deep learning methods DOI Creative Commons
Muthu Rama Krishnan Mookiah, Roberto Puch‐Solis,

Niamh Nic Daéid

и другие.

Forensic Science International, Год журнала: 2023, Номер 349, С. 111734 - 111734

Опубликована: Май 19, 2023

Ballistics (the linkage of bullets and cartridge cases to weapons) is a common type evidence encountered in criminal around the world. The interest lies determining whether two were fired using same firearm. This paper proposes an automated method classify from surface topography Land Engraved Area (LEA) images pellets machine deep learning methods. curvature was removed loess fit features extracted Empirical Mode Decomposition (EMD) followed by various entropy measures. informative identified minimum Redundancy Maximum Relevance (mRMR), finally classification performed Support Vector Machines (SVM), Decision Tree (DT) Random Forest (RF) classifiers. results revealed good predictive performance. In addition, model DenseNet121 used LEA images. provided higher performance than SVM, DT RF Moreover, Grad-CAM technique visualise discriminative regions These suggest that proposed can be expedite projectiles firearms assist ballistic examinations. this work, compared air both rifles high velocity pistol. Air guns collect data because they more accessible other could as proxy, delivering comparable LEAs. methods developed here proof-of-concept are easily expandable bullet case identification any weapon.

Язык: Английский

Процитировано

2
A Bayesian approach based on Kalman filter frameworks for bullet identification DOI
Hamed Danandeh Hesar,

Saeed Bigdeli,

Mohsen Ebrahimi Moghaddam

и другие.

Science & Justice, Год журнала: 2019, Номер 59(4), С. 390 - 404

Опубликована: Фев. 26, 2019

Язык: Английский

Процитировано

3
A Multimodal Fusion Approach for Bullet Identification Systems DOI

Saeed Bigdeli,

Mohsen Ebrahimi Moghaddam

Journal of Forensic Sciences, Год журнала: 2018, Номер 64(3), С. 741 - 753

Опубликована: Ноя. 21, 2018

Abstract In the field of forensic science, bullet identification is based on fact that firing cartridge from a barrel leaves exclusive microscopic striation fired bullets as fingerprint firearm. The methods are categorized in 2‐D and 3‐D their image acquisition techniques. this study, we focus optical images using multimodal technique propose several distinct its modalities. proposed method uses rule‐based linear weighted fusion approach which combines semantic level decisions different modalities with optimized weights have been identified by genetic algorithm. was applied dataset, includes 180 90 AK ‐47 barrels. experimentations showed our attained better results compared to common identification.

Язык: Английский

Процитировано

1
Fault Diagnosis Based on Tree Heuristic Feature Selection and FS-DFV for Rolling Element Bearings DOI Open Access

Xiaoyue chen,

Xiong Liu, Ge Dang

и другие.

IOP Conference Series Materials Science and Engineering, Год журнала: 2019, Номер 630(1), С. 012024 - 012024

Опубликована: Окт. 1, 2019

Abstract In order to make up for the deficiency of traditional single diagnosis in rolling element bearing fault application, eliminate a large amount redundant information and improve classification effect aliasing mode, based on comprehensive analysis respective advantages fuzzy set tree search, this paper presents joint method tree-inspired feature selection FS-DFV (Fuzzy Set Dependent Feature Vector). The dependent vectors (DFV) can dig deeper essential differences faults accuracy. By establishing heuristic model, type search strategy is designed, excellent criteria density clustering with noise are proposed, conventional model improved. addition, used process problem extracting patterns DFV, membership guide subsequent extraction alias modes. proposed compared other four methods. experimental results show that effectively diagnostic efficiency bearing.

Язык: Английский

Процитировано

0