Multi-Magnification Attention Convolutional Neural Networks [AI-eXplained]

Chia Wei Chao; Daniel Winden Hwang; Hung-Wen Tsai; Shih Hsuan Lin; Wei Li Chen; Chun Rong Huang; Pau Choo Chung

doi:10.1109/MCI.2023.3277771

Multi-Magnification Attention Convolutional Neural Networks [AI-eXplained]

Chia Wei Chao, Daniel Winden Hwang, Hung-Wen Tsai, Shih Hsuan Lin, Wei Li Chen, Chun Rong Huang, Pau Choo Chung

Research output: Contribution to journal › Article › peer-review

2 Citations (Scopus)

Abstract

To apply convolutional neural networks (CNNs) on high-resolution images, a common approach is to split the input image into smaller patches. However, the field-of-view is restricted by the input size. To overcome the problem, a multi-magnification attention convolutional neural network (MMA-CNN) is proposed to analyze images based on both local and global features. Our approach focuses on identifying the importance of individual features at each magnification level and is applied to pathology whole slide images (WSIs) segmentation to show its effectiveness. Several interactive figures are also developed to enhance the reader's understanding of our research.

Original language	English
Pages (from-to)	54-55
Number of pages	2
Journal	IEEE Computational Intelligence Magazine
Volume	18
Issue number	3
DOIs	https://doi.org/10.1109/MCI.2023.3277771
Publication status	Published - 2023 Aug 1

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Artificial Intelligence

Access to Document

10.1109/MCI.2023.3277771

Cite this

@article{cb84879e9d04415dadf68f965cac6f49,

title = "Multi-Magnification Attention Convolutional Neural Networks [AI-eXplained]",

abstract = "To apply convolutional neural networks (CNNs) on high-resolution images, a common approach is to split the input image into smaller patches. However, the field-of-view is restricted by the input size. To overcome the problem, a multi-magnification attention convolutional neural network (MMA-CNN) is proposed to analyze images based on both local and global features. Our approach focuses on identifying the importance of individual features at each magnification level and is applied to pathology whole slide images (WSIs) segmentation to show its effectiveness. Several interactive figures are also developed to enhance the reader's understanding of our research.",

author = "Chao, {Chia Wei} and Hwang, {Daniel Winden} and Hung-Wen Tsai and Lin, {Shih Hsuan} and Chen, {Wei Li} and Huang, {Chun Rong} and Chung, {Pau Choo}",

note = "Funding Information: This work was supported in part by the National Science and Technology Council, Taiwan under Grants NSTC 111-2634-F-006-012 and NSTC 111- 2628-E-006-011-MY3. We would like to thank National Center for High-performance Computing (NCHC) for providing computational and storage resources. Publisher Copyright: {\textcopyright} 2005-2012 IEEE.",

year = "2023",

month = aug,

day = "1",

doi = "10.1109/MCI.2023.3277771",

language = "English",

volume = "18",

pages = "54--55",

journal = "IEEE Computational Intelligence Magazine",

issn = "1556-603X",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "3",

}

TY - JOUR

T1 - Multi-Magnification Attention Convolutional Neural Networks [AI-eXplained]

AU - Chao, Chia Wei

AU - Hwang, Daniel Winden

AU - Tsai, Hung-Wen

AU - Lin, Shih Hsuan

AU - Chen, Wei Li

AU - Huang, Chun Rong

AU - Chung, Pau Choo

N1 - Funding Information: This work was supported in part by the National Science and Technology Council, Taiwan under Grants NSTC 111-2634-F-006-012 and NSTC 111- 2628-E-006-011-MY3. We would like to thank National Center for High-performance Computing (NCHC) for providing computational and storage resources. Publisher Copyright: © 2005-2012 IEEE.

PY - 2023/8/1

Y1 - 2023/8/1

N2 - To apply convolutional neural networks (CNNs) on high-resolution images, a common approach is to split the input image into smaller patches. However, the field-of-view is restricted by the input size. To overcome the problem, a multi-magnification attention convolutional neural network (MMA-CNN) is proposed to analyze images based on both local and global features. Our approach focuses on identifying the importance of individual features at each magnification level and is applied to pathology whole slide images (WSIs) segmentation to show its effectiveness. Several interactive figures are also developed to enhance the reader's understanding of our research.

AB - To apply convolutional neural networks (CNNs) on high-resolution images, a common approach is to split the input image into smaller patches. However, the field-of-view is restricted by the input size. To overcome the problem, a multi-magnification attention convolutional neural network (MMA-CNN) is proposed to analyze images based on both local and global features. Our approach focuses on identifying the importance of individual features at each magnification level and is applied to pathology whole slide images (WSIs) segmentation to show its effectiveness. Several interactive figures are also developed to enhance the reader's understanding of our research.

UR - http://www.scopus.com/inward/record.url?scp=85165910210&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85165910210&partnerID=8YFLogxK

U2 - 10.1109/MCI.2023.3277771

DO - 10.1109/MCI.2023.3277771

M3 - Article

AN - SCOPUS:85165910210

SN - 1556-603X

VL - 18

SP - 54

EP - 55

JO - IEEE Computational Intelligence Magazine

JF - IEEE Computational Intelligence Magazine

IS - 3

ER -

Multi-Magnification Attention Convolutional Neural Networks [AI-eXplained]

Abstract

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this