A novel model of feature extraction for lung cysts detection in CT image using Minutiae based Mumford and Shah functional model

A Novel Diagnostic Model For Lung Cancer Detection using Mumford-Shah and SVM Classifier

“…The extracted features such as area, circumference, centroid, diameter, eccentricity, and mean intensity of pixels were used to train the SVM method and a trained model is created. The training time for the classification is 4.3 seconds which is an improvement over the existing method [10].…”

“…The circularity, posterior dimension, area, eccentricity and shape features are extracted to train the SVM and also to classify in order to identify the benign or malignant nodule. Kishore Sebastian and S. Devi [10] developed a system based on Mumford and Shah model for diagnosing lung cancer by analyzing CT images of the lung. The system includes filtering, image segmentation, binarization and thinning, minimization of false minute points and finally the GLCM is for feature extraction.…”

“…The work does not include the classification of ROI using the available types of classifiers. The proposed system uses the same method [10] and classifies cancer nodule as benign or malignant using a SVM. Also the results obtained using SVM classifier is compared with KNN, Decision-Tree, and Adaptive Boosting algorithms.…”

“…Each image is 512 x 512 pixels in size as an unsigned integer. Modifications have been made to the existing solution [10], and the proposed model is shown in Fig. 1.…”

“…However, some of the research works does not show any satisfactory accuracy in the detection of the disease and the works still need to be improved to grab high accuracy closer to 100%. Image processing methods [10] and machine learning methods have been implemented in order to detect and also to classify the type of lung cancer. The latest systems so far developed for lung cancer detection based on CT scan images are thoroughly reviewed, analyzed and proposed a new model.…”