Abstract: In this paper, we have discussed the background estimation using “3 parameter based running average method”, and then used background subtraction followed by an appropriate thresholding operation to detect the moving object. Our main aim is to do the background modeling to get moving object. In this process, detection of moving object directly depends on background model. Numerical experiments confirm our theoretical results.

• Publication Date: 28-Apr-2013
• DOI: 10.15224/978-981-07-6260-5-51
• Views: 0
• Downloads: 0

Abstract: To improve the performance of the content based image retrieval (CBIR), relevance feedback (RF) plays an important role as it allows interactive image retrieval. Different RF techniques have been proposed, such as density estimation RF algorithm, subspace learning algorithm, classification based algorithm etc. The existing density estimation algorithm considers only positive samples and classification based algorithms considers positive and negative samples in different groups . Subspace learning algorithm assumes that positive and negative samples are linearly distributed .But existing algorithms are having the main drawback of semantic gap as well as ‘small sample size’ problem. In this paper, we discuss various RF techniques proposed earlier in the literature. The advantages of using a graph embedding framework for RF are also discussed. In addition this paper also provides a comparative study of various methods proposed by researchers in RF.

• Publication Date: 28-Apr-2013
• DOI: 10.15224/978-981-07-6260-5-52
• Views: 0
• Downloads: 0