建國科技大學圖書館 |

語系: 繁體中文

說明(常見問題)

圖書館個人資料蒐集告知聲明

登入

回首頁

切換: 標籤 | MARC模式 | ISBD

Extreme value theory-based methods f...

Scheirer, Walter J.

Extreme value theory-based methods for visual recognition

紀錄類型:	書目-語言資料,印刷品 : 單行本
作者:	ScheirerWalter J.,
出版地:	[San Rafael, California]
出版者:	Morgan & Claypool Publishers;
出版年:	2017.
面頁冊數:	xv, 115 p.ill. : 24 cm.;
集叢名:	Synthesis lectures on computer vision#10
標題:	Computer vision. -
標題:	Extreme value theory. -
標題:	Visual perception. -
摘要註:	A common feature of many approaches to modeling sensory statistics is an emphasis on capturing the "average." From early representations in the brain, to highly abstracted class categories in machine learning for classification tasks, central-tendency models based on the Gaussian distribution are a seemingly natural and obvious choice for modeling sensory data. However, insights from neuroscience, psychology, and computer vision suggest an alternate strategy: preferentially focusing representational resources on the extremes of the distribution ofsensory inputs. The notion of treating extrema near a decision boundary as features is not necessarily new, but a comprehensive statistical theory of recognition based on extrema is only now just emerging in the computer vision literature. This book beginsby introducing the statistical Extreme Value Theory (EVT) for visual recognition. In contrast to central-tendency modeling, it is hypothesized that distributions near decision boundaries form a more powerful model for recognition tasks by focusing coding resources on data that are arguably the most diagnostic features.EVT has several important properties: strong statistical grounding, better modeling accuracy near decision boundaries thanGaussian modeling, the ability to model asymmetric decision boundaries, and accurate prediction of the probability of an event beyond our experience. The second part of the book uses thetheory to describe a new class of machine learning algorithms fordecision making that are a measurable advance beyond the state-of-the-art. This includes methods for post-recognition score analysis, information fusion, multi-attribute spaces, and calibration of supervised machine learning algorithms.
ISBN:	9781627057004

Extreme value theory-based methods for visual recognition
Scheirer, Walter J.

Extreme value theory-based methods for visual recognition / Walter J. Scheirer. - [San Rafael, California] : Morgan & Claypool Publishers, 2017.. - xv, 115 p. ; ill. ; 24 cm.. - (Synthesis lectures on computer vision ; #10).
Includes bibliographical references (p. 99-114).
ISBN 9781627057004ISBN 1627057005ISBN 162705703X
Computer vision.Extreme value theory.Visual perception.

Extreme value theory-based methods for visual recognition
LDR:02600nam0a2200241 4500 001 378770
010 1 $a 9781627057004 $b pbk. $d NT1512
010 1 $a 1627057005 $b pbk.
010 1 $a 162705703X $b ebook
100 $a 20180528d2017 y0engy01 b
101 0 $a eng
102 $a us $b ca
105 $a a a 000yy
200 1 $a Extreme value theory-based methods for visual recognition $f Walter J. Scheirer.
210 $a [San Rafael, California] $c Morgan & Claypool Publishers $d 2017.
215 1 $a xv, 115 p. $c ill. $d 24 cm.
225 2 $a Synthesis lectures on computer vision $v #10 $x 2153-1064
320 $a Includes bibliographical references (p. 99-114)
330 $a A common feature of many approaches to modeling sensory statistics is an emphasis on capturing the "average." From early representations in the brain, to highly abstracted class categories in machine learning for classification tasks, central-tendency models based on the Gaussian distribution are a seemingly natural and obvious choice for modeling sensory data. However, insights from neuroscience, psychology, and computer vision suggest an alternate strategy: preferentially focusing representational resources on the extremes of the distribution ofsensory inputs. The notion of treating extrema near a decision boundary as features is not necessarily new, but a comprehensive statistical theory of recognition based on extrema is only now just emerging in the computer vision literature. This book beginsby introducing the statistical Extreme Value Theory (EVT) for visual recognition. In contrast to central-tendency modeling, it is hypothesized that distributions near decision boundaries form a more powerful model for recognition tasks by focusing coding resources on data that are arguably the most diagnostic features.EVT has several important properties: strong statistical grounding, better modeling accuracy near decision boundaries thanGaussian modeling, the ability to model asymmetric decision boundaries, and accurate prediction of the probability of an event beyond our experience. The second part of the book uses thetheory to describe a new class of machine learning algorithms fordecision making that are a measurable advance beyond the state-of-the-art. This includes methods for post-recognition score analysis, information fusion, multi-attribute spaces, and calibration of supervised machine learning algorithms.
410 0 $1 2001 $a Synthesis lectures on computer vision $v #10 $1 0111 $a 2153-1056.
410 0 $1 2001 $a Synthesis lectures on computer vision $v #10 $x 2153-1064
606 $a Computer vision. $2 lc $3 27610
606 $a Extreme value theory. $2 lc $3 366135
606 $a Visual perception. $2 lc $3 39569
676 $a 006.37 $v 23
680 $a TA1634 $b .S354 2017
700 1 $a Scheirer $b Walter J. $3 366134
801 0 $a cw $b CTU $c 20180529 $g AACR2