Abstract | ||
---|---|---|
The similarity measure for image pairs plays a predominant role in image registration. Generally, mutual information (MI) or normalized mutual information (NMI), been defined by the density functions, is often adopted as the similarity measure in image registration. In this paper, based on the proposed survival entropy (SE), a new similarity measure, refer to as the cross survival entropy (CSE), is introduced by using the cumulative distributions. As a new and more generalized form of similarity measure, comparing with MI and cross-cumulative residual entropy (CCRE), we elucidate some excellent properties of CSE. Numerous contrastive implements have shown that CSE achieves more robustness and more accuracy in image registration, which confirm the validity of SE and CSE. |
Year | DOI | Venue |
---|---|---|
2011 | 10.1109/ICIS.2011.35 | ACIS-ICIS |
Keywords | Field | DocType |
cross survival entropy,cross-cumulative residual entropy,image registration,proposed survival entropy,image pair,normalized mutual information,mutual information,new similarity measure,similarity measure,cumulative distribution,entropy,random variables,density functional theory,distribution functions,accuracy,computed tomography | Cross entropy,Random variable,Pattern recognition,Similarity measure,Joint entropy,Mutual information,Differential entropy,Artificial intelligence,Residual entropy,Image registration,Mathematics | Conference |
Citations | PageRank | References |
1 | 0.41 | 5 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Shiwei Yu | 1 | 68 | 9.54 |
Xiaoyun Liu | 2 | 49 | 7.10 |
Wufan Chen | 3 | 511 | 59.06 |