Bag-of-visual words based automatic image annotation

摘要: Content-based Image retrieval systems extract and retrieve images using their low-level features, such as color, texture, shape. Nevertheless, these visual contents do not allow a user to formulate semantically meaningful image query. annotation are solution solve the inadequacy of CBIR text based retrieval. There have been several studies on automatic utilizing machine learning techniques images' representation with low level features extracted either global or local methods. However, typically, approaches suffer from correlation between globally assigned annotations used obtain automatically. In this paper, we present an approach enhance effectiveness bag word that is created automatically set manually annotated training images. The experimentation performed 4,000 for training, 1000 testing ImageNet. result has shown 77.5% performance accuracy. work believed be one step towards enhancing existing minimizing semantic gap.