Abstract
The medical field has its significance with increasing the demand of automatic diagnosis. These automated systems reduce the effort of the experts to make decisions. Our proposed system supports experts making the right decisions while predicting the cancer tumors in the lungs based on the CT image scan. This system converts RGB images into gray images, removes the noise using the median filter, and segments the CT images to avoid the unwanted part from the scanned image because of the segmented images’ discriminative features. Those features are extracted by using the Local Binary Patterns. Finally, the classification was done by the SVM kernels, such as linear, polynomial, and radial basis function. The radial basis kernel function achieved 88.76% accuracy. The proposed approach is tested on the LIDC dataset.
