國立虎尾科技大學 |

Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation.

Record Type:	Language materials, printed : Monograph/item
Title/Author:	Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation./
Author:	Cheng, Zeyuan.
Published:	Ann Arbor : ProQuest Dissertations & Theses, : 2022,
Description:	59 p.
Notes:	Source: Masters Abstracts International, Volume: 83-12.
Contained By:	Masters Abstracts International83-12.
Subject:	Engineering. -
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28967543
ISBN:	9798834061397

Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation.
Cheng, Zeyuan.

Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation. - Ann Arbor : ProQuest Dissertations & Theses, 2022 - 59 p.

Source: Masters Abstracts International, Volume: 83-12.

Thesis (M.A.S.)--University of Toronto (Canada), 2022.

This item must not be sold to any third party vendors.

The increasing demand for computational power is pushing the photolithography technology to its physical limit. As one of its crucial technical bottlenecks, the design issue of layout hotspots requires innovative solutions to minimize the necessity of conducting expensive lithography simulation experiments but remain extremely accurate. In addition to the sustained development of conventional hotspots detection, the recent implementation of convolutional neural networks (CNN) has shown high effectiveness in flexible pattern recognitions. However, limited by the industrial data availability, technical issues such as imbalanced learning and rare/new hotspot detection need to be navigated. In this thesis, we introduced a deep learning hotspots detection method with an improved CNN classifier and enhanced the learning effectiveness by augmenting the dataset with the generative adversarial network (GAN) based synthetic pattern generation. Experiment results show competitive performance compared to the existing works, but further development is required to connect with the current industrial processes.

ISBN: 9798834061397Subjects--Topical Terms:

561152
Engineering.
Subjects--Index Terms:

Deep learning

Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation.
LDR:02315nam a2200373 4500 001 1104576
005 20230619080053.5
006 m o d
007 cr#unu||||||||
008 230907s2022 ||||||||||||||||| ||eng d
020 $a 9798834061397
035 $a (MiAaPQ)AAI28967543
035 $a AAI28967543
040 $a MiAaPQ $c MiAaPQ
100 1 $a Cheng, Zeyuan. $0 (orcid)0000-0003-1504-0710 $3 1413436
245 1 0 $a Accurate Hotspots Detection with Deep Convolutional Neural Network and Generative Adversarial Network Based Data Augmentation.
260 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2022
300 $a 59 p.
500 $a Source: Masters Abstracts International, Volume: 83-12.
500 $a Advisor: Behdinan, Kamran.
502 $a Thesis (M.A.S.)--University of Toronto (Canada), 2022.
506 $a This item must not be sold to any third party vendors.
520 $a The increasing demand for computational power is pushing the photolithography technology to its physical limit. As one of its crucial technical bottlenecks, the design issue of layout hotspots requires innovative solutions to minimize the necessity of conducting expensive lithography simulation experiments but remain extremely accurate. In addition to the sustained development of conventional hotspots detection, the recent implementation of convolutional neural networks (CNN) has shown high effectiveness in flexible pattern recognitions. However, limited by the industrial data availability, technical issues such as imbalanced learning and rare/new hotspot detection need to be navigated. In this thesis, we introduced a deep learning hotspots detection method with an improved CNN classifier and enhanced the learning effectiveness by augmenting the dataset with the generative adversarial network (GAN) based synthetic pattern generation. Experiment results show competitive performance compared to the existing works, but further development is required to connect with the current industrial processes.
590 $a School code: 0779.
650 4 $a Engineering. $3 561152
650 4 $a Electrical engineering. $3 596380
650 4 $a Mechanical engineering. $3 557493
653 $a Deep learning
653 $a Generative adversarial network
653 $a Hotspot detection
690 $a 0537
690 $a 0544
690 $a 0548
710 2 $a University of Toronto (Canada). $b Mechanical and Industrial Engineering. $3 845502
773 0 $t Masters Abstracts International $g 83-12.
790 $a 0779
791 $a M.A.S.
792 $a 2022
793 $a English
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=28967543