Computerized Lung Nodule Detection Using 3D Feature Extraction and Learning Based Algorithms

Ozekes S., Osman O.

JOURNAL OF MEDICAL SYSTEMS, cilt.34, sa.2, ss.185-194, 2010 (SCI-Expanded) identifier identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 34 Sayı: 2
  • Basım Tarihi: 2010
  • Doi Numarası: 10.1007/s10916-008-9230-0
  • Dergi Adı: JOURNAL OF MEDICAL SYSTEMS
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus
  • Sayfa Sayıları: ss.185-194
  • Anahtar Kelimeler: 3D feature extraction, Feed-forward neural networks, Support vector machines, Naive Bayes, Logistic regression
  • Marmara Üniversitesi Adresli: Hayır

Özet

In this paper, a Computer Aided Detection (CAD) system based on three-dimensional (3D) feature extraction is introduced to detect lung nodules. First, eight directional search was applied in order to extract regions of interests (ROIs). Then, 3D feature extraction was performed which includes 3D connected component labeling, straightness calculation, thickness calculation, determining the middle slice, vertical and horizontal widths calculation, regularity calculation, and calculation of vertical and horizontal black pixel ratios. To make a decision for each ROI, feed forward neural networks (NN), support vector machines (SVM), na < ve Bayes (NB) and logistic regression (LR) methods were used. These methods were trained and tested via k-fold cross validation, and results were compared. To test the performance of the proposed system, 11 cases, which were taken from Lung Image Database Consortium (LIDC) dataset, were used. ROC curves were given for all methods and 100% detection sensitivity was reached except na < ve Bayes.