A Neural Network Model for Automatic Image Annotation Refinement
Alpesh Dabhi, Bhavesh Prajapati · Journal of Emerging Technologies and Innovative Research · 2014
In image retrieval techniques, automatic image annotation (AIA) is used to label image by set of semantic keywords automatically. Due to its potential impact on semantic based image retrieval, AIA has been an active research topic in recent year. AIA uses machine learning algorithm for classification of network, this network is use to automatically classify image content. However, the results generated by classification network are still far from satisfaction because it is very difficult to map low level visual features of image to high level semantic feature. Thus refinement of automatic image annotation is necessary. The purpose of refining image annotation is to reserve highly correlated annotations and remove weakly or irrelevant annotation. In this paper, a novel neural network based method is proposed to refine automatic image annotation. In this method correlation between keywords are identified by using neural network and final refined result can generate based on this correlation of keyword.