A novel neural network model for shrimp segmentation to detect white spot syndrome

Image segmentation is an essential part of almost any image processing methodology and it play a critical role in protecting the region of interest on any substrate image before its actual analysis is prescribed. In fact, the accuracy of any processing done by image segmentation will largely depends on the efficiency of the segmentation algorithm employed. A typical segmentation method employing a important features of Canny-GLCM (Gray Level Co-occurrence Matrix) incorporated with a simple Artificial Neural Network (ANN) model is proposed in this research work for segmentation of shrimp variability. Performance metrics related to accuracy have been compared with benchmark of this method, and the sensitivity of efficiency level has been described. The segmentation in the proposed research work is targeted towards Penaeus Monodon (PM), and Litopenaeus Vannamei (LV) diversities for main threats detection of White Spot Syndrome (WSS). The proposed model has better performance metrics, such as (94.67%), sensitivity (94.79%), specificity (94.51%) and positive predictive (94.79%) while compared to other existing methods.