Abstract

Abstract: As weare living in multimedia era, nowadays Internet has been heavily used forviewing, sharing and storing collection of the videos and images. This resultsin huge amount of data. In user's point of view, different domains are facingmany problems with retrieving images that are relevant to their query. To solvethese problems different techniques are adopted.  Significant work has been done in severalissues of performance evaluation such as feature extraction, feature matching,semantic gap reduction and similarity measurements. Based on the requirement ofdifferent applications and current technologies, we retrieve, process and storethe content based images. But, we are lacking techniques to process, retrieveand store images faster. The Hadoop Map Reduce platform provides a system forlarge storage and computationally intensive distributed processing when, it isused for Content-based Image Retrieval (CBIR) systems performance can beimproved compared to other retrieval systems.

Keywords: Annotation,Content-based Image Retrieval, Feature extraction, Relevance feedback,Similarity measure, Computer Vision, Hadoop HDFS, MapReduce Framework.