Analysis of Supervised and Unsupervised Learning in Content Based Multimedia Retrieval

摘要

This work analyze images using different learning algorithms such as supervised and unsupervised algorithm for image retrieval. The color and shape based image descriptors are used to distinguish specific set of images from the whole database. Wang's database is taken to extract the ten clustered color and shape features such as ten-dominant color descriptor and Zernike Moments from the image. The supervised learning algorithms named as Support Vector Machine (SVM) K-Nearest Neighbour (KNN), Naive-Bayes and Random Forest are trained by this feature vectors are used for class prediction. Subsequently, highest probability class images are grouped to reduce the search space for similar images. The distance measure is used to calculate the similarity between the user query and newly grouped images.