An investigation of CNN models for differentiating malignant from benign lesions using small pathologically proven datasets

An investigation of CNN models for differentiating malignant from benign lesions using small pathologically proven datasets

Highlights•Proposed two CNN models to classify small cancer image dataset (Malignance/Benign). •Combine raw images and LBP features can improve the classification on small data. •Proposed V-1D model can better study the small and unbalanced dataset. •Local information from lung nodule significantly...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Computerized medical imaging and graphics 2019-10, Vol.77, p.101645-101645, Article 101645
Hauptverfasser: Zhang, Shu, Han, Fangfang, Liang, Zhengrong, Tan, Jiaxing, Cao, Weiguo, Gao, Yongfeng, Pomeroy, Marc, Ng, Kenneth, Hou, Wei
Format: Artikel
Sprache:eng
Schlagworte:
Algorithms
Artificial neural networks
Benign
Cancer
Cancer imaging
Classification
Computed tomography
Convolutional neural network
Datasets
Ground truth
Image classification
Internal Medicine
Learning algorithms
Lesions
Lung nodules
Machine learning
Medical imaging
Neural networks
Nodule characterization
Nodules
Other
Pathologically proven datasets
Polyp characterization
Polyps
Radiology
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Highlights•Proposed two CNN models to classify small cancer image dataset (Malignance/Benign). •Combine raw images and LBP features can improve the classification on small data. •Proposed V-1D model can better study the small and unbalanced dataset. •Local information from lung nodule significantly improves the M/B classification.
ISSN:0895-6111
1879-0771
DOI:10.1016/j.compmedimag.2019.101645