國立虎尾科技大學 |

A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions./
作者:	Whang, Esther.
面頁冊數:	1 online resource (126 pages)
附註:	Source: Masters Abstracts International, Volume: 84-11.
Contained By:	Masters Abstracts International84-11.
標題:	Medical imaging. -
電子資源:	click for full text (PQDT)
ISBN:	9798379507527

A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions.
Whang, Esther.

A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions. - 1 online resource (126 pages)

Source: Masters Abstracts International, Volume: 84-11.

Thesis (M.E.)--The Cooper Union for the Advancement of Science and Art, 2023.

Includes bibliographical references

Developments in whole slide imaging and in computational analysis methods have opened the door to new discoveries in the field of medicine. In particular, machine learning has gained attention as a tool for analyzing large amounts of medical image data. However, while the ease of using machine learning systems makes them appealing, building the labeled datasets required to train these systems remains a difficult problem due to the need for trained professionals to annotate each data point. This work presents a machine learning-based classification pipeline that incorporates both labeled and unlabeled images, reducing the burden of building a fully-labeled dataset. The pipeline consists of a pretrained convolutional neural network feature extractor and a support vector machine classifier, with the feature extractor finetuned with MixMatch, a semi-supervised learning algorithm. Different training methods are explored based on the pretrained weights of the feature extractor and the ratio of labeled to unlabeled images used during training. The final classification pipelines are evaluated on a heart tissue dataset from The Human Protein Atlas that was annotated for the presence of intercalated discs by doctors at Johns Hopkins University. This approach achieves higher accuracy in classifying intercalated disc presence than a supervised learning-based pipeline.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798379507527Subjects--Topical Terms:

1180167
Medical imaging.
Subjects--Index Terms:

Deep learningIndex Terms--Genre/Form:

554714
Electronic books.

A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions.
LDR:02825ntm a22003977 4500 001 1142794
005 20240429063835.5
006 m o d
007 cr mn ---uuuuu
008 250605s2023 xx obm 000 0 eng d
020 $a 9798379507527
035 $a (MiAaPQ)AAI30490741
035 $a AAI30490741
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Whang, Esther. $3 1467240
245 1 2 $a A Semi-Supervised Learning Approach to Classifying Pathological Images of Heart Tissue for the Presence of Intercellular Junctions.
264 0 $c 2023
300 $a 1 online resource (126 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: 84-11.
500 $a Advisor: Fontaine, Fred L.
502 $a Thesis (M.E.)--The Cooper Union for the Advancement of Science and Art, 2023.
504 $a Includes bibliographical references
520 $a Developments in whole slide imaging and in computational analysis methods have opened the door to new discoveries in the field of medicine. In particular, machine learning has gained attention as a tool for analyzing large amounts of medical image data. However, while the ease of using machine learning systems makes them appealing, building the labeled datasets required to train these systems remains a difficult problem due to the need for trained professionals to annotate each data point. This work presents a machine learning-based classification pipeline that incorporates both labeled and unlabeled images, reducing the burden of building a fully-labeled dataset. The pipeline consists of a pretrained convolutional neural network feature extractor and a support vector machine classifier, with the feature extractor finetuned with MixMatch, a semi-supervised learning algorithm. Different training methods are explored based on the pretrained weights of the feature extractor and the ratio of labeled to unlabeled images used during training. The final classification pipelines are evaluated on a heart tissue dataset from The Human Protein Atlas that was annotated for the presence of intercalated discs by doctors at Johns Hopkins University. This approach achieves higher accuracy in classifying intercalated disc presence than a supervised learning-based pipeline.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Medical imaging. $3 1180167
650 4 $a Biomedical engineering. $3 588770
653 $a Deep learning
653 $a Digital pathology
653 $a Machine learning
653 $a Semi-supervised learning
653 $a Whole slide imaging
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0574
690 $a 0541
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a The Cooper Union for the Advancement of Science and Art. $b Electrical Engineering. $3 1467241
773 0 $t Masters Abstracts International $g 84-11.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=30490741 $z click for full text (PQDT)