Underwater sound target identification model optimization method based on sample expansion network

• Main Authors: CHEN WEIDONG, LUO HENGGUANG, WANG DAYU, LI JIN, ZHANG BOXUAN, ZHAO TIANBAI, WANG SHAOBO
• Format: Patent
• Language: Chinese; English
• Published: 04.11.2022
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

The invention provides an underwater acoustic target recognition model optimization method based on a sample expansion network, and belongs to the field of underwater acoustic target recognition model optimization. In order to solve the problem of low classification accuracy caused by insufficient target data, two pairs of generators and discriminators with completely symmetrical structures are established by utilizing a mask-based sample generation thought on the premise of ensuring the real-time effect of a model, and a source domain sample is mapped to a target domain. Experimental results show that the cyclic generative adversarial network is built through the mask prompt thought, while the clear model structure is ensured, the reliable and real sample of the target domain is generated, the training set is added to optimize the recognition model, and the recognition accuracy is improved. 本发明提出了一种基于样本扩充网络的水声目标识别模型优化方法，属于水声目标识别模型优化领域。本发明针对目标数据不足导致的分类准确率不高的问题，利用基于掩模的样本生成思想，在保证模型实时效果的前提下，搭建了两对结构完全对称的生成器和判别器，将