Automatic Image Labeling System for Outdoor Scene

Abstract

Automatic image annotation is important to both video and image retrieval and computer object recognition. To ensure easy sharing of an effective searching over a huge and fast growing number of online images, real-time automatic annotation by words is an imperative but highly challenging task. The system tends to partition the image plane to a set of non-overlapping regions with coherent image appearance. In the proposed system, the efficient segmentation algorithm is developed by combining the N-cut segmentation algorithm to partition the image region and color component analysis to extract center region. In the system, there are six number of segmented parts are used. Five segmented parts are retrieved by using N-cut and center image is retrieved using color component analysis method.