Evolution of ubiquitous computing in the areas of personal computing technology has produced staggeringly large data It is difficult to search mainly the image data by understanding users objective only by keywords and phrases and this leads to uncertain outcomes. For producing these outcomes effectively, this paper introduces a new approach to the problem of image learning to enable search engines to learn about visual content over time based on user feedback through one click activity and images from a pool recovered by text based query are re-ranked depending on both visual and text based query. Content Based Image Retrieval (CBIR) techniques are used for accessing semantically-relevant images from an image data source depending on automatically-derived image functions for features like Geometric moments, Global histogram, Color Moments, Local histogram. Documents can also be retrieved using the text based query by the user.
展开▼
