Underwater Backscatter Recognition Using Deep Fuzzy Extreme Convolutional Neural Network Optimized via Hunger Games Search

Although deep learning methods are accurate in underwater backscatter detection, identification, and classification, they suffer from long processing times, especially in the training phase. Therefore, a four-phase deep learning (DL) based approach is proposed for real-time underwater backscatter classification. First, a deep convolutional neural network (DCNN) is exploited as a feature extraction section. Secondly, in order to reduce training and testing time, the extreme learning machine (ELM) substitutes the fully connected layer. Using ELM in the last layer causes uncertainty and unreliability; therefore, in the third stage, the hunger games search (HGS) will be used to tackle this shortcoming. Finally, fuzzy systems are used to balance the relationship between the HGS’s exploration and exploitation phases. For evaluating the efficiency of the designed fuzzy HGS (FHGS), we first use 23 standard benchmark mathematical optimization functions. Subsequently, we employ three experimental sonar datasets to examine the efficiency of DCNN-ELM-FHGS in dealing with high-dimensional datasets. For a comprehensive investigation, we compare FHGS to the standard HGS, Whale Optimization Algorithm, Gray Wolf Optimizer, Kalman Filter, Henry Gas Solubility Optimization, Harris Hawks Optimization, Chimp Optimization Algorithm, Genetic Algorithm, and Particle Swarm Optimization, with respect to convergence rate, entrapment in local minima, and detection accuracy. The results demonstrate that the proposed strategy performs better in detecting underwater anomaly targets by an average of 2.11 percent compared to the best benchmark model.