Segmentation and Classification of Ct Cervix Images Using Bag of Visual Word Classifier

Abstract

Image processing is used in the medical field for detection of tumor. Image segmentation is a vital part of image processing. Segmentation is the process of partitioning an image into distinct regions. The algorithm has the steps of preprocessing, cervix extraction, cervix boundary correction, image segmentation, feature extraction and image classification. The image is preprocessed using Adaptive median filtering and Fuzzy thresholding. The cervix is extracted using canny edge detection and border tracing algorithm. The cervix boundary correction is performed using Adaptive Concave Hull algorithm. Segmentation is performed using Region growing based technique. Then for the segmented tumor region, the features are extracted using the GLCM (Gray Level Co-occurrence Matrix) algorithm. From the features extracted, the image is classified as the benign or malignant cervix by using the BOVW(Bag of Visual Word) classifier.