Saliency based shape extraction of objects in unconstrained underwater environment

摘要: Un-manned underwater exploration in unconstrained environment is a challenging and non-trivial problem. Manual analysis of large volume images/videos captured by the stations/vehicles major bottleneck for research community. Automated system analyzing these videos need hour exploring space. In this paper, we present method extracting shape objects scenarios. The proposed extracts using saliency gradient based morphological active contour models. uniqueness that stopping condition models derived from combination with scene. As result able to work highly dynamic environments. results show extract shapes man-made as well natural environmental conditions. detect multiple an successful occluded such GAC minimum 63% average 90% misclassification rate 4% whereas ACWE 62% 85% 4%.