Enhanced Semantic Image Retrieval using Feature Extraction and KNN Techniques

Abstract

In addition of that, the technique required some additional techniques to correct the retrieval process such as user feedback, these methods consumes additional time of search. Thus a new technique with hybrid concept is proposed for improving the content based image search. The proposed technique includes the technique to train the system using the image features and text for annotation of image. For identifying the images more accurately the text and image features are used. Finally to retrieve the data (image) using user query (image or text) a KNN algorithm is implemented with it. The implementation of the proposed model is performed using visual studio technology and their performance in terms of time and space complexity is estimated. In addition of that the performance in terms of accuracy and error rate is also provided for demonstrating the relevancy of image search. General Terms Your general terms must be any term which can be used for general classification of the submitted material such as Pattern Recognition, Security, Algorithms et. al.