東華大學圖書館 |

語系: 繁體中文

說明(常見問題)

回圖書館首頁

手機版館藏查詢

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

FindBook

Google Book

Amazon

博客來

Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification.

紀錄類型:	書目-電子資源 : Monograph/item
正題名/作者:	Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification./
作者:	Bharadwaj, Akshay Ganesh.
面頁冊數:	1 online resource (73 pages)
附註:	Source: Masters Abstracts International, Volume: 81-11.
Contained By:	Masters Abstracts International81-11.
標題:	Industrial engineering. -
電子資源:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27820127click for full text (PQDT)
ISBN:	9781658440103

Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification.
Bharadwaj, Akshay Ganesh.

Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification. - 1 online resource (73 pages)

Source: Masters Abstracts International, Volume: 81-11.

Thesis (M.Sc.)--North Carolina State University, 2020.

Includes bibliographical references

Contemporary research in a multitude of engineering disciplines is focused towards leveraging the massive amount of data that has been generated over the last 2 decades. Efficient utilization of this data is possible through the use of machine learning techniques, specifically, deep learning methods that utilize artificial neural networks to automatically infer patterns from data. In the field of product design and manufacturing, there is a large amount of valuable data being generated both in industry as well as in academic settings. Several robust systems have been developed for recording end-point manufacturing machine data. However, there is a lack of easy access to large amounts of diverse Computer-Aided Design (CAD) data, to train such algorithms and methods to classify and search through this data. In this work, proof of concept for a design and manufacturing cyber-infrastructure (CI) entitled 'FabWave', is demonstrated. The aim of the CI tool is to provide an inter-operable platform for easy sharing and access of product design and manufacturing data, which is also able to accommodate a variety of design data formats and inputs from various design systems. In the pilot implementation, the design data is captured through automated workflows from product design classes at two different universities, along with data aggregated from multiple online stores of product design data. The data is processed in order to separate individual parts from assemblies and obtain related part metadata. These are then stored in standard STEP format, select platformspecific formats such as .F3D, and tessellated file formats, in order to enable multiple methods of local feature or part-class detection. Based on the data thus obtained, a method for classification of product design data is proposed, building on a state-of-art multi-view Convolutional Neural Network method used in computer graphics. This new network is termed MVCNN++. This method involves capture of multiple images of each part, which are provided as input to the network. A classification scheme is proposed for the data based on relaxed part-type designation scheme, in order to supply data during the training process. The addition of each part's associated dimension data is shown to yield a superior classification accuracy, with an improvement of 6% on average in comparison to the state-of-art computer graphics methods. A method for evaluation of this Neural Network classification task on the unclassified data corpus is described, along with a discussion of the results. This work lays the foundation of a cyber-infrastructure necessary to connect and collect design data from academia with potential new design paradigms in making available manufacturing data generated from within academia and from public sources.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2023

Mode of access: World Wide Web

ISBN: 9781658440103Subjects--Topical Terms:

526216
Industrial engineering.
Subjects--Index Terms:

Cyber-infrastructureIndex Terms--Genre/Form:

542853
Electronic books.

Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification.
LDR:04263nmm a2200385K 4500 001 2358038
005 20230725094932.5
006 m o d
007 cr mn ---uuuuu
008 241011s2020 xx obm 000 0 eng d
020 $a 9781658440103
035 $a (MiAaPQ)AAI27820127
035 $a (MiAaPQ)NCState_Univ18402037211
035 $a AAI27820127
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Bharadwaj, Akshay Ganesh. $3 3698570
245 1 0 $a Inception of a Cyber-Infrastructure for Product Design Data and Evaluation of a Customized Multi-View Convolutional Neural Network for 3D CAD Model Classification.
264 0 $c 2020
300 $a 1 online resource (73 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 81-11.
500 $a Advisor: Starly, Binil.
502 $a Thesis (M.Sc.)--North Carolina State University, 2020.
504 $a Includes bibliographical references
520 $a Contemporary research in a multitude of engineering disciplines is focused towards leveraging the massive amount of data that has been generated over the last 2 decades. Efficient utilization of this data is possible through the use of machine learning techniques, specifically, deep learning methods that utilize artificial neural networks to automatically infer patterns from data. In the field of product design and manufacturing, there is a large amount of valuable data being generated both in industry as well as in academic settings. Several robust systems have been developed for recording end-point manufacturing machine data. However, there is a lack of easy access to large amounts of diverse Computer-Aided Design (CAD) data, to train such algorithms and methods to classify and search through this data. In this work, proof of concept for a design and manufacturing cyber-infrastructure (CI) entitled 'FabWave', is demonstrated. The aim of the CI tool is to provide an inter-operable platform for easy sharing and access of product design and manufacturing data, which is also able to accommodate a variety of design data formats and inputs from various design systems. In the pilot implementation, the design data is captured through automated workflows from product design classes at two different universities, along with data aggregated from multiple online stores of product design data. The data is processed in order to separate individual parts from assemblies and obtain related part metadata. These are then stored in standard STEP format, select platformspecific formats such as .F3D, and tessellated file formats, in order to enable multiple methods of local feature or part-class detection. Based on the data thus obtained, a method for classification of product design data is proposed, building on a state-of-art multi-view Convolutional Neural Network method used in computer graphics. This new network is termed MVCNN++. This method involves capture of multiple images of each part, which are provided as input to the network. A classification scheme is proposed for the data based on relaxed part-type designation scheme, in order to supply data during the training process. The addition of each part's associated dimension data is shown to yield a superior classification accuracy, with an improvement of 6% on average in comparison to the state-of-art computer graphics methods. A method for evaluation of this Neural Network classification task on the unclassified data corpus is described, along with a discussion of the results. This work lays the foundation of a cyber-infrastructure necessary to connect and collect design data from academia with potential new design paradigms in making available manufacturing data generated from within academia and from public sources.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2023
538 $a Mode of access: World Wide Web
650 4 $a Industrial engineering. $3 526216
650 4 $a Artificial intelligence. $3 516317
653 $a Cyber-infrastructure
653 $a Product design data
653 $a Neural network
653 $a 3D CAD model classification
655 7 $a Electronic books. $2 lcsh $3 542853
690 $a 0800
690 $a 0546
710 2 $a ProQuest Information and Learning Co. $3 783688
710 2 $a North Carolina State University. $3 1018772
773 0 $t Masters Abstracts International $g 81-11.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=27820127 $z click for full text (PQDT)