SHI Hong1, ZHAO Chun-hui1, SHEN Zheng-yan2, et al. Sonar image retrieval system based on maximum weighted region mutual information[J]. 2011, 30(10): 102-105.
SHI Hong1, ZHAO Chun-hui1, SHEN Zheng-yan2, et al. Sonar image retrieval system based on maximum weighted region mutual information[J]. 2011, 30(10): 102-105. DOI: 10.13873/j.1000-97872011.10.025.
a sonar image retrieval system based on maximum weighted region mutual information(MI)is designed.Sonar image segmentation is carried out using spectrum segmentation method.The region MI value between the example image and every image in the image database is calculated
and weighted region MI is obtained since that sonar image retrieval is point at the sonar target objects.The maximum weighted region MI is used instead of MI
and image sorting according to the size order of it is carried out
the images is taken arranged in the front as the retrieval result by selecting certain threshold value.The simulation result indicates that the maximum weighted region MI method is more suitable for the sonar image retrieval especially in dealing with two images with same background and different foreground or same foreground and different background.