Multi-Scale Fish Segmentation Refinement Using Contour Based Segmentation

Image processing and the analysis techniques are the increasing attention when they have enabled the non-extractive and the non-lethal approach for the collection of fisheries data. The data collection includes the following requirements such as fish size, catch estimation, regulatory compliance, species recognition and population counting. The main process that is used to measure the size of fish accurately is image segmentation. The challenges that can affect the segmentation of images include the blurring of the image areas due to the water droplets on the camera lens and the fish bodies which are out of the camera view. This project describes the automatic segmentation of fish for underwater images This segmentation algorithm implemented for identify the shape of the fish contour-based segmentation is implemented in this project. The project describes about the issues with an effective contour-based segmentation from an initial segmentation. The refinement is processed from coarse level to fine level. At the coarse level, the entire fish is aligned for the contour of the initial segmentation with trained representative contours by using iteratively reweighted least squares (IRLS). At finer levels, the refinement of contour segments is done to represent poorly segmented or missing shape parts. This method addresses the problems listed above and generates promising results with highly robust segmentation performance and length measurement.