Automatic object detection and segmentation from underwater images via saliency-based region merging
Yafei Zhu, Lin Chang, Jialun Dai, Haiyong Zheng, Bing Zheng · OCEANS 2016 - Shanghai · 2016
Underwater object detection and segmentation has been attracting a lot of interest, and recently various systems have been designed. In this paper, we introduce a novel technique to automatically detect and segment objects from underwater images via saliency-based region merging. The method is composed of three main steps. Firstly, a salient object detection model is used to detect the position of salient objects in underwater image. Secondly, background prior is applied to determine the approximate background location. Thirdly, the region merging based interactive image segmentation method is improved by adding the determined object and background location information as the user inputs so that the algorithm becomes automatic. The experimental results show that it's efficient to segment objects from the underwater image by the proposed method.