Open AccessUsing Distance Measure based Classification in Automatic Extraction of Lungs Cancer Nodules for Computer Aided Diagnosis
Author(s) -
Maan Ammar,
Muhammad Shamdeen,
Mazen Kasedeh,
Kinan Mansour,
Waad Ammar
Publication year - 2021
Publication title -
signal and image processing : an international journal
Language(s) - English
Resource type - Journals
eISSN - 2229-3922
pISSN - 0976-710X
DOI - 10.5121/sipij.2021.12303
Subject(s) - nodule (geology) , computer science , euclidean distance , artificial intelligence , pattern recognition (psychology) , feature extraction , computer aided diagnosis , extraction (chemistry) , measure (data warehouse) , lung cancer , computer vision , radiology , medicine , pathology , data mining , biology , paleontology , chemistry , chromatography
We introduce in this paper a reliable method for automatic extraction of lungs nodules from CT chest images and shed the light on the details of using the Weighted Euclidean Distance (WED) for classifying lungs connected components into nodule and not-nodule. We explain also using Connected Component Labeling (CCL) in an effective and flexible method for extraction of lungs area from chest CT images with a wide variety of shapes and sizes. This lungs extraction method makes use of, as well as CCL, some morphological operations. Our tests have shown that the performance of the introduce method is high. Finally, in order to check whether the method works correctly or not for healthy and patient CT images, we tested the method by some images of healthy persons and demonstrated that the overall performance of the method is satisfactory.