Image segmentation and classification based on CNN model to detect brain tumor
No Thumbnail Available
Date
2022
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Open Access Color
OpenAIRE Downloads
OpenAIRE Views
Abstract
Tipik olarak, bir beyin tümörünün ciddiyetini sınıflandırmak ve belirlemek için son beyin ameliyatından önce biyopsi yapılmaz. Makine öğrenimi ve yapay zeka gibi daha az müdahaleci tümör tespit teknolojilerinin gelecekte erişilebilir hale gelmesi bekleniyor. CNN algoritması, resimleri (CNN) segmentlere ayırma ve sınıflandırmada olağanüstü performans gösteren bir makine öğrenme tekniğidir. Beyin tümörü segmentasyonu ve sınıflandırması için aşağıdaki mimari, burada tartışma için önerilmektedir. Üç farklı tümör modalitesine dayanmaktadır. Yeni oluşturulan ağı, önceki sistemlerden çok daha temel olan T1 kontrastlı manyetik rezonans görüntüleme kullanarak analiz etmek için seçildi. Ağın genellenebilirliği, daha büyük bir görüntü veri tabanı ve (on) konuya özgü çapraz doğrulama tekniklerinden biri kullanılarak, aramada kullanılacak şekilde değerlendirildi. 10 katlı çapraz doğrulama tekniği, kayıt odaklı çalışmada en iyi sonucu verdi. yüzde 96,56 doğrulukla daha büyük veri kümesinin çapraz doğrulaması, onu en doğru yöntem haline getiriyor. Büyük genelleme kapasitesi ve kısa yanıt süresi ile yeni geliştirilen CNN mimarisi, tıbbi tanısal radyologlar için harika bir karar destek aracı olabilir.
Typically, no biopsy is performed prior to final brain surgery in order to classify and establish the seriousness of a brain tumor. It is anticipated that less intrusive tumor detection technologies, such as machine learning and artificial intelligence , will become accessible in the future. The CNN algorithm is a machine learning technique that has shown remarkable performance in segmenting and classifying pictures (CNN). The following architecture for brain tumor segmentation and classification is proposed for discussion here. It is based on three distinct tumor modalities. It was chosen to analyze the newly formed network using T1 contrast-enhanced magnetic resonance imaging, which is far more basic than the previous systems. The network's generalizability was assessed using a larger image database and (one of the ten) topic-specific cross-validation techniques that will be used in the search.. The 10-fold cross-validation technique produced the best result in the record-oriented cross-validation of the larger dataset, with an accuracy of 96.56 percent, making it the most accurate method. With its great generalization capacity and short response time, the newly developed CNN architecture may show to be an amazing decision support tool for medical diagnostic radiologists..
Typically, no biopsy is performed prior to final brain surgery in order to classify and establish the seriousness of a brain tumor. It is anticipated that less intrusive tumor detection technologies, such as machine learning and artificial intelligence , will become accessible in the future. The CNN algorithm is a machine learning technique that has shown remarkable performance in segmenting and classifying pictures (CNN). The following architecture for brain tumor segmentation and classification is proposed for discussion here. It is based on three distinct tumor modalities. It was chosen to analyze the newly formed network using T1 contrast-enhanced magnetic resonance imaging, which is far more basic than the previous systems. The network's generalizability was assessed using a larger image database and (one of the ten) topic-specific cross-validation techniques that will be used in the search.. The 10-fold cross-validation technique produced the best result in the record-oriented cross-validation of the larger dataset, with an accuracy of 96.56 percent, making it the most accurate method. With its great generalization capacity and short response time, the newly developed CNN architecture may show to be an amazing decision support tool for medical diagnostic radiologists..
Description
Keywords
Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Computer Engineering and Computer Science and Control
Turkish CoHE Thesis Center URL
Fields of Science
Citation
WoS Q
Scopus Q
Source
Volume
Issue
Start Page
End Page
88