Image Search Re-ranking based on Prototype

Till now whatever the methods are available for image search are based on the text provided in the search. but such kind of image search re-ranking suffer from unreliability because the resulting images contain more irrelevant images that of the serach. Hence the re-ranking concept arises to re-rank the retrieved images based on the text surrounding the image and metadata and visual feature. In this paper, we propose an “Image Search Re-ranking Based on Prototype” a prototype based method to learn reranking function from human labeled samples . Based on images obtained in initial search result , visual prototype will be generated. Each prototype is used to construct a meta reranker to produce a reranking score for any other image from initial set. Finally all scores from all meta rerankers are aggregated. For visual reranking we use SVM algorithm.