Content Based Medical Image Retrieval System using SIFT and Hu-Moment for Hepatobiliary Images

摘要

This study presents retrieval of hepatobiliary images based on content-based medical image retrieval technique. Hepatobiliary images are related to hepatobiliary diseases. Hepatobiliary diseases include a group of diseases related to liver and biliary system. For image retrieval, low-level feature is used. For feature extraction sift (scale invariant feature transform) and Hu-moments methods are used. Features extracted by these methods are combining to form a single feature vector. Feature extracted by query images and images which are stored in the database are compared using Euclidean distance method. Based on minimum distance images are retrieved. For performance measurement precision, recall and error rate are measured.